AI Impact on Jobs, Workers, and Wages

Score ↕	Year ↕	Title	Authors ↕	Theme ↕
10	2023	Generative AI at Work core ↗ This paper directly addresses the core project themes by empirically measuring the causal effects of generative AI on worker productivity and analyzing distributional impacts across skill levels. It specifically investigates the mechanism of AI augmenting labor for novice workers and its implications for the traditional career ladder and entry-level performance. We study the staggered introduction of a generative AI-based conversational assistant using data from 5,179 customer support agents.Access to the tool increases productivity, as measured by issues resolved per hour, by 14 percent on average, with the greatest impact on novice and lowskilled workers, and minimal impact on experienced and highly skilled workers.We provide suggestive evidence that the AI model disseminates the potentially tacit knowledge of more able workers and helps newer workers move down the experience curve.In addition, we show that AI assistance improves customer sentiment, reduces requests for managerial intervention, and improves employee retention.	Erik Brynjolfsson, Danielle Li, Lindsey Raymond
10	2024	The simple macroeconomics of AI core ↗ This paper directly addresses the project's core questions on macroeconomic effects, distributional consequences, and task-based measurement of AI. It explicitly analyzes aggregate employment and wage impacts, skill complementarity versus substitution, and the resulting inequality, aligning perfectly with the researcher's focus. SUMMARY This paper evaluates claims about the large macroeconomic implications of new advances in Artificial intelligence (AI). It starts from a task-based model of AI’s effects, working through automation and task complementarities. So long as AI’s microeconomic effects are driven by cost savings/productivity improvements at the task level, its macroeconomic consequences will be given by a version of Hulten’s theorem: Gross Domestic Product (GDP) and aggregate productivity gains can be estimated by what fraction of tasks are impacted and average task-level cost savings. Using existing estimates on exposure to AI and productivity improvements at the task level, these macroeconomic effects...	Daron Acemoğlu
10	2025	Artificial Intelligence and the Labor Market core ↗ This paper directly addresses the core project questions by constructing novel measures of AI exposure and analyzing its causal effects on labor demand and employment. It provides critical insights into whether AI acts as a substitute or complement to labor and examines the resulting aggregate labor market outcomes, aligning perfectly with the study's focus on task-based frameworks and distributional effects. We use advances in natural language processing to construct new measures of workers' task-level exposure to artificial intelligence (AI) and machine learning from 2010 to 2023, capturing variation across firms, occupations, and time.Tasks with higher AI exposure subsequently experience reduced labor demand.To interpret these patterns, we develop a model that separates direct substitution from indirect reallocative effects of labor-saving technologies.Two variables summarize the impact of AI on within-firm labor demand: the mean exposure of an occupation's tasks, which depresses demand, and the concentration of exposure in a few tasks, which offsets losses by enabling workers to reallocate...	Menaka Hampole, Dimitris Papanikolaou, Lawrence Schmidt, et al.
9	2023	Navigating the Jagged Technological Frontier: Field Experimental Evidence of the Effects of AI on Knowledge Worker Productivity and Quality core ↗ This paper directly addresses the project's core question of whether AI tools augment or substitute for labor by providing rigorous experimental evidence on AI's heterogeneous effects on knowledge worker productivity and quality. It empirically illustrates the 'jagged' nature of AI adoption, showing significant gains for tasks within the technological frontier while highlighting performance declines for complex tasks outside it, which is central to understanding distributional effects and task reorganization. We introduce and study the concept of a “jagged technology frontier” to describe the uneven impact of artificial intelligence (AI) capabilities, where AI assistance improves performance for some tasks but worsens it for others, even within the same knowledge workflow and with a seemingly similar level of difficulty. In collaboration with the global management consulting firm Boston Consulting Group, we have developed realistic management consulting tasks and examined the human performance implications of using AI to perform complex and knowledge-intensive work. The preregistered experiment involved 758 knowledge workers. After establishing a performance baseline on similar tasks, subjects...	Fabrizio Dell’Acqua, Edward McFowland, Ethan Mollick, et al.
9	2023	GPTs are GPTs: An Early Look at the Labor Market Impact Potential of Large Language Models core ↗ This paper directly addresses the core theme of measuring AI exposure across occupations by developing a rubric to assess alignment with LLM capabilities. It provides essential empirical context on the potential productivity impacts and task automation risks, which are central to understanding how AI reshapes labor markets. We investigate the potential implications of large language models (LLMs), such as Generative Pre-trained Transformers (GPTs), on the U.S. labor market, focusing on the increased capabilities arising from LLM-powered software compared to LLMs on their own. Using a new rubric, we assess occupations based on their alignment with LLM capabilities, integrating both human expertise and GPT-4 classifications. Our findings reveal that around 80% of the U.S. workforce could have at least 10% of their work tasks affected by the introduction of LLMs, while approximately 19% of workers may see at least 50% of their tasks impacted. We do not make predictions about the development or adoption timeline...	Tyna Eloundou, Sam Manning, Pamela Mishkin, et al.
9	2024	New Frontiers: The Origins and Content of New Work, 1940–2018 core ↗ This paper directly addresses the project's core question of whether technology augments or substitutes for labor by analyzing the long-term evolution of 'new work' in response to automation and augmentation innovations. It provides essential historical context and empirical evidence on how technological shocks reshape occupational demand, which is crucial for understanding the current AI-driven labor market transformation. Abstract We answer three core questions about the hypothesized role of newly emerging job categories (“new work”) in counterbalancing the erosive effect of task-displacing automation on labor demand: what is the substantive content of new work, where does it come from, and what effect does it have on labor demand? We construct a novel database spanning eight decades of new job titles linked to U.S. Census microdata and to patent-based measures of occupations’ exposure to labor-augmenting and labor-automating innovations. The majority of current employment is in new job specialties introduced since 1940, but the locus of new-work creation has shifted from middle-paid production and clerical...	David Autor, Caroline Chin, Anna Salomons, et al.
9	2020	AI and Jobs: Evidence from Online Vacancies core ↗ This paper directly addresses the project's core question regarding how AI reshapes labor markets by providing granular evidence on hiring patterns and AI exposure using a large-scale online vacancy dataset. It specifically investigates the substitution versus augmentation dynamics and aggregate effects, offering key empirical insights into whether AI is currently displacing workers or merely changing task composition. We study the impact of AI on labor markets using establishment-level data on vacancies with detailed occupation and skill information comprising the near-universe of online vacancies in the US from 2010 onwards. There is rapid growth in AI related vacancies over 2010-2018 that is greater in AI-exposed establishments. AI-exposed establishments are reducing hiring in non-AI positions. We find no discernible relationship between AI exposure and employment or wage growth at the occupation or industry level, however, implying that AI is currently substituting for humans in a subset of tasks but it is not yet having detectable aggregate labor market consequences.	Daron Acemoğlu, David Autor, Jonathon Hazell, et al.
9	2025	Still Waters, Rapid Currents: Early Labor Market Transformation under Generative AI core ↗ This paper directly addresses the project's core questions regarding the causal effects of generative AI on worker productivity, wages, and employment by leveraging high-quality matched employer-employee data. It provides critical aggregate evidence on whether AI tools currently augment or substitute labor, finding minimal immediate labor market transformation despite widespread adoption. We examine the labor market effects of AI chatbots using two large-scale adoption surveys (late 2023 and 2024) covering 11 exposed occupations (25,000 workers, 7,000 workplaces), linked to matched employer-employee data in Denmark.AI chatbots are now widespread-most employers encourage their use, many deploy in-house models, and training initiatives are common.These firm-led investments boost adoption, narrow demographic gaps in take-up, enhance workplace utility, and create new job tasks.Yet, despite substantial investments, economic impacts remain minimal.Using difference-in-differences and employer policies as quasi-experimental variation, we estimate precise zeros: AI chatbots have had...	Anders Humlum, Emilie Vestergaard
9	2025	Making Talk Cheap: Generative AI and Labor Market Signaling core ↗ [Title only] This title directly addresses the labor market implications of generative AI, specifically focusing on signaling mechanisms which is a core theme of the project. It likely explores how AI affects worker productivity and inequality by altering the cost of communication, fitting squarely within the distributional effects and task-based framework interests. No abstract available.	Anaïs Galdin, Jesse Silbert
9	2025	Technology and Labor Markets: Past, Present, and Future; Evidence from Two Centuries of Innovation core ↗ This paper directly addresses the project's core themes by developing novel AI exposure measures and analyzing their causal effects on labor market distribution across occupations and demographics. It specifically tackles the question of whether AI acts as a substitute or complement for labor and identifies potential winners and losers, providing critical historical context and forward-looking predictions relevant to the research scope. We use recent advances in natural language processing and large language models to construct novel measures of technology exposure for workers that span almost two centuries.Combining our measures with Census data on occupation employment, we show that technological progress over the 20th century has led to economically meaningful shifts in labor demand across occupations: it has consistently increased demand for occupations with higher education requirements, occupations that pay higher wages, and occupations with a greater fraction of female workers.Using these insights and a calibrated model, we then explore different scenarios for how advances in artificial intelligence (AI) are likely...	Huben Liu, Dimitris Papanikolaou, Lawrence Schmidt, et al.
9	2025	Job Transformation, Specialization, and the Labor Market Effects of AI core ↗ This paper directly addresses the project's core questions by modeling how AI transforms job task content and affecting wage inequality and labor market dynamics. It provides specific insights into skill complementarity versus substitution and the distributional effects of generative AI on workers, which are central to the researcher's investigation. A central effect of automation is to transform jobs-shifting their task content. We develop a general-equilibrium model of this process. Occupations bundle tasks; workers possess task-specific skills and sort by comparative advantage. When a task is automated, remaining tasks gain in importance, so wage effects depend on workers' full skill profiles. We estimate the distribution of task-specific skills and project individual-level wage effects of generative-AI automation. Moderate exposure benefits workers on average but high exposure harms them, with large dispersion within occupations; the return to social skills rises, that to analytical skills falls; and low-earners gain more than...	Lukas B. Freund, Lukas Friedrich Mann
9	2017	The Future of Employment: How Susceptible Are Jobs to Computerisation? core [Title only] This seminal paper by Frey and Osborne is foundational for the project's core theme of measuring AI exposure across occupations by estimating the probability of computerisation. It directly informs the task-based framework and distributional effects analysis by identifying which jobs are susceptible to automation, serving as a key reference for understanding the causal mechanisms of AI on labor markets. No abstract available.	Frey, Osborne
9	2023	Experimental Evidence on the Productivity Effects of Generative AI core [Title only] This title directly addresses the core theme of generative AI productivity experiments, which is a primary research question for the project. It suggests empirical findings on how AI impacts worker output, allowing for analysis of augmentation versus substitution effects. No abstract available.	Noy, Zhang
9	2025	Canaries in the Coal Mine? Six Facts about the Recent Employment Effects of Artificial Intelligence core [Title only] The title explicitly addresses recent employment effects of AI, directly tackling the project's core questions regarding causal effects on employment and aggregate labor market trends. The metaphor 'Canaries in the Coal Mine' suggests an analysis of early warning signs or specific occupational impacts, which aligns with investigating distributional effects and identifying winners and losers across different worker groups. No abstract available.	Brynjolfsson, Chandar, Chen
9	2024	Displacement or Complementarity? The Labor Market Impact of Generative AI core [Title only] The title directly addresses the core dichotomy of substitution versus augmentation central to the researcher's inquiry into how AI reshapes labor markets. It explicitly focuses on generative AI, which aligns perfectly with the project's interest in measuring AI exposure and its distributional effects on workers. No abstract available.	Chen, Srinivasan, Zakerinia
8	2019	Automation and New Tasks: How Technology Displaces and Reinstates Labor core ↗ This paper establishes the foundational task-based framework that underpins the project's core themes of labor displacement, reinstatement, and the measurement of technology exposure. It provides essential theoretical context for analyzing how automation and AI reshape labor markets through specific mechanisms of task allocation and worker substitution. We present a framework for understanding the effects of automation and other types of technological changes on labor demand, and use it to interpret changes in US employment over the recent past. At the center of our framework is the allocation of tasks to capital and labor—the task content of production. Automation, which enables capital to replace labor in tasks it was previously engaged in, shifts the task content of production against labor because of a displacement effect. As a result, automation always reduces the labor share in value added and may reduce labor demand even as it raises productivity. The effects of automation are counterbalanced by the creation of new tasks in which...	Daron Acemoğlu, Pascual Restrepo
8	2021	Tasks, Automation, and the Rise in US Wage Inequality core ↗ This paper establishes the foundational task-based framework and empirical evidence regarding how automation displaces routine tasks, which is central to the project's theoretical and measurement approaches. It provides critical historical context on how technological shifts reshape labor markets, offering a benchmark for analyzing the distinct impacts of modern AI on wages and inequality. We document that between 50% and 70% of changes in the US wage structure over the last four decades are accounted for by the relative wage declines of worker groups specialized in routine tasks in industries experiencing rapid automation. We develop a conceptual framework where tasks across a number of industries are allocated to different types of labor and capital. Automation technologies expand the set of tasks performed by capital, displacing certain worker groups from employment opportunities for which they have comparative advantage. This framework yields a simple equation linking wage changes of a demographic group to the task displacement it experiences. We report robust evidence in...	Daron Acemoğlu, Pascual Restrepo
8	2021	Measuring the Occupational Impact of AI: Tasks, Cognitive Abilities and AI Benchmarks core ↗ This paper provides a novel methodological framework for measuring AI exposure by linking occupational tasks to cognitive abilities and AI benchmarks, directly addressing the project's core theme of AI exposure measurement. It offers valuable insights into which specific cognitive abilities and occupations are most susceptible to AI, informing the discussion on who the winners and losers are based on skill levels and task types. In this paper we develop a framework for analysing the impact of Artificial Intelligence (AI) on occupations. This framework maps 59 generic tasks from worker surveys and an occupational database to 14 cognitive abilities (that we extract from the cognitive science literature) and these to a comprehensive list of 328 AI benchmarks used to evaluate research intensity across a broad range of different AI areas. The use of cognitive abilities as an intermediate layer, instead of mapping work tasks to AI benchmarks directly, allows for an identification of potential AI exposure for tasks for which AI applications have not been explicitly created. An application of our framework to occupational...	Songül Tolan, Annarosa Pesole, Fernando Martínez‐Plumed, et al.
8	2025	Large Language Models, Small Labor Market Effects core ↗ [Title only] The title directly addresses the project's core interest in the causal effects of generative AI and large language models on labor markets. It likely provides aggregate evidence on employment or wage impacts, which is a key question regarding whether AI tools are substituting for or augmenting labor. No abstract available.	Anders Humlum, Emilie Vestergaard
8	2025	The (Short-Term) Effects of Large Language Models on Unemployment and Earnings core ↗ This paper directly addresses the project's core question regarding the causal effects of AI on employment and wages by empirically testing the impact of Large Language Models on earnings and unemployment. It provides valuable evidence on whether AI initially acts as a complement or substitute for labor, specifically finding that exposure leads to earnings increases without immediate job displacement. Large Language Models have spread rapidly since the release of ChatGPT in late 2022, accompanied by claims of major productivity gains but also concerns about job displacement. This paper examines the short-run labor market effects of LLM adoption by comparing earnings and unemployment across occupations with differing levels of exposure to these technologies. Using a Synthetic Difference in Differences approach, we estimate the impact of LLM exposure on earnings and unemployment. Our findings show that workers in highly exposed occupations experienced earnings increases following ChatGPT's introduction, while unemployment rates remained unchanged. These results suggest that initial labor...	Danqing Chen, Carina Kane, Austin C. Kozlowski, et al.
8	2026	AI Skills Improve Job Prospects: Causal Evidence from a Hiring Experiment core ↗ This paper directly addresses the project's core theme of how AI skills affect labor market outcomes by providing causal evidence on their value in hiring decisions. It specifically explores the distributional effects of AI exposure, examining how these skills mitigate disadvantages for older or less educated workers and vary across different occupations. The growing adoption of artificial intelligence (AI) technologies has heightened interest in the labor market value of AI related skills, yet causal evidence on their role in hiring decisions remains scarce. This study examines whether AI skills serve as a positive hiring signal and whether they can offset conventional disadvantages such as older age or lower formal education. We conducted an experimental survey with 1,725 recruiters from the United Kingdom, the United States and Germany. Using a paired conjoint design, recruiters evaluated hypothetical candidates represented by synthetically designed resumes. Across three occupations of graphic design, office assistance, and software...	Fabian Stephany, Ole Teutloff, Angelo Leone
8	2017	What Can Machines Learn, and What Does It Mean for Occupations and the Economy? core [Title only] The title directly addresses the core project themes by linking machine learning capabilities to occupational and economic impacts, suggesting a focus on AI exposure measurement and task-based frameworks. It likely explores the theoretical underpinnings of AI substitution versus augmentation across different worker skill levels, which is central to the research questions. No abstract available.	Brynjolfsson, Mitchell
8	2024	Is Generative AI a Job Killer? Evidence from the Online Freelance Market core [Title only] This paper directly addresses the core themes of AI substitution versus augmentation and distributional effects within the specific context of online labor markets. It provides empirical evidence on whether generative AI displaces workers, which is central to understanding the causal effects of AI on employment and the future of work. No abstract available.	Hui, Reshef, Zhou
7	2003	The Skill Content of Recent Technological Change: An Empirical Exploration core ↗ This paper establishes the foundational task-based framework that defines how technology substitutes for routine tasks and complements non-routine ones, a core mechanism for understanding AI's impact. While it focuses on historical computerization rather than modern generative AI, its theoretical structure is essential for analyzing whether current AI tools act as substitutes or complements across different skill levels and occupations. We apply an understanding of what computers do to study how computerization alters job skill demands. We argue that computer capital (1) substitutes for workers in performing cognitive and manual tasks that can be accomplished by following explicit rules; and (2) complements workers in performing nonroutine problem-solving and complex communications tasks. Provided that these tasks are imperfect substitutes, our model implies measurable changes in the composition of job tasks, which we explore using representative data on task input for 1960 to 1998. We find that within industries, occupations, and education groups, computerization is associated with reduced labor input of routine manual...	David Autor, Frank Levy, Richard J. Murnane
7	2024	Applying AI to Rebuild Middle Class Jobs core ↗ This paper closely aligns with the project's interest in AI skill complementarity and the potential for AI to augment labor rather than substitute it, specifically regarding middle-class occupations. However, as a theoretical argument rather than an empirical study, it lacks the causal evidence and quantitative measurement of productivity or wage effects central to the research questions. While the utopian vision of the current Information Age was that computerization would flatten economic hierarchies by democratizing information, the opposite has occurred.Information, it turns out, is merely an input into a more consequential economic function, decision-making, which is the province of elite experts.The unique opportunity that AI offers to the labor market is to extend the relevance, reach, and value of human expertise.Because of AI's capacity to weave information and rules with acquired experience to support decision-making, it can be applied to enable a larger set of workers possessing complementary knowledge to perform some of the higher-stakes decision-making tasks...	David Autor
7	2018	Replication data for: The Race between Man and Machine: Implications of Technology for Growth, Factor Shares, and Employment core ↗ This paper provides the foundational task-based theoretical framework essential for analyzing how technology substitutes for or complements labor, a core theme of the project. It establishes the mechanisms for automation and new task creation that underpin empirical assessments of AI's impact on employment, wages, and inequality. We examine the concerns that new technologies will render labor redundant in a framework in which tasks previously performed by labor can be automated and new versions of existing tasks, in which labor has a comparative advantage, can be created. In a static version where capital is fixed and technology is exogenous, automation reduces employment and the labor share, and may even reduce wages, while the creation of new tasks has the opposite effects. Our full model endogenizes capital accumulation and the direction of research toward automation and the creation of new tasks. If the long-run rental rate of capital relative to the wage is sufficiently low, the long-run equilibrium involves...	Acemoglu, Daron, Restrepo, Pascual
6	2018	Is Automation Labor-Displacing? Productivity Growth, Employment, and the Labor Share core [Title only] This paper addresses the core macroeconomic question of AI's impact on employment and labor share, though it focuses on traditional automation rather than modern generative AI. It is highly relevant to the aggregate labor market and distributional effects themes but may lack specific insights into the current wave of machine learning and LLM adoption. No abstract available.	Autor, Salomons
5	2019	The Impact of Artificial Intelligence on the Labor Market core ↗ [Title only] The title is overly broad and generic, making it impossible to determine if the paper addresses specific core themes like task-based frameworks, generative AI, or distributional effects. It could range from a high-level macroeconomic review to a narrow empirical study, leading to significant uncertainty about its precise relevance to the project. No abstract available.	Michael A. Webb	Other
5	2015	Why Are There Still So Many Jobs? The History and Future of Workplace Automation core [Title only] This title suggests a broad historical and theoretical discussion on workplace automation, which likely addresses long-standing debates rather than focusing specifically on modern AI technologies like machine learning or large language models. While it may touch on the persistence of jobs, it lacks the specific empirical or causal focus on recent AI impacts on labor markets that defines the core research project. No abstract available.	Autor
10	2023	Experimental evidence on the productivity effects of generative artificial intelligence ↗ This paper directly addresses the project's core theme of generative AI productivity experiments by providing causal evidence on how AI tools augment worker performance in writing tasks. It specifically answers questions regarding the magnitude of productivity gains, the impact on output quality, and the immediate effects on inequality, which are central to understanding the labor market impacts of AI. We examined the productivity effects of a generative artificial intelligence (AI) technology, the assistive chatbot ChatGPT, in the context of midlevel professional writing tasks. In a preregistered online experiment, we assigned occupation-specific, incentivized writing tasks to 453 college-educated professionals and randomly exposed half of them to ChatGPT. Our results show that ChatGPT substantially raised productivity: The average time taken decreased by 40% and output quality rose by 18%. Inequality between workers decreased, and concern and excitement about AI temporarily rose. Workers exposed to ChatGPT during the experiment were 2 times as likely to report using it in their real job...	Shakked Noy, Whitney Zhang
10	2025	Generative AI at Work ↗ This paper directly addresses the project's core question on whether AI tools augment or substitute for labor by providing causal evidence that generative AI significantly boosts customer support worker productivity. It also contributes to the theme of distributional effects and entry-level workers by demonstrating substantial gains for less experienced and lower-skilled agents, alongside improvements in work experience and learning. Abstract We study the staggered introduction of a generative AI–based conversational assistant using data from 5,172 customer-support agents. Access to AI assistance increases worker productivity, as measured by issues resolved per hour, by 15% on average, with substantial heterogeneity across workers. The effects vary significantly across different agents. Less experienced and lower-skilled workers improve both the speed and quality of their output, while the most experienced and highest-skilled workers see small gains in speed and small declines in quality. We also find evidence that AI assistance facilitates worker learning and improves English fluency, particularly among international...	Erik Brynjolfsson, Danielle Li, Lindsey Raymond
10	2023	Experimental Evidence on the Productivity Effects of Generative Artificial Intelligence ↗ [Title only] This title directly addresses the core theme of generative AI productivity experiments, which is central to understanding the causal effects of AI on worker performance. It explicitly focuses on the mechanism of augmentation versus substitution, a key question in determining how AI reshapes labor markets and wages. No abstract available.	Shakked Noy, Whitney Zhang
10	2024	The Simple Macroeconomics of AI ↗ This paper directly addresses the project's core questions regarding the aggregate macroeconomic effects of AI, specifically focusing on TFP, employment, and wage inequality. It utilizes a task-based framework to analyze how AI exposure translates into productivity gains and distributional outcomes, aligning perfectly with the researcher's interest in macroeconomic impacts and skill complementarity. SUMMARY This paper evaluates claims about the large macroeconomic implications of new advances in Artificial intelligence (AI). It starts from a task-based model of AI’s effects, working through automation and task complementarities. So long as AI’s microeconomic effects are driven by cost savings/productivity improvements at the task level, its macroeconomic consequences will be given by a version of Hulten’s theorem: Gross Domestic Product (GDP) and aggregate productivity gains can be estimated by what fraction of tasks are impacted and average task-level cost savings. Using existing estimates on exposure to AI and productivity improvements at the task level, these macroeconomic effects...	Daron Acemoğlu
10	2023	Occupational Heterogeneity in Exposure to Generative AI ↗ [Title only] The title directly addresses the core theme of measuring AI exposure across different occupations, which is a foundational question for the project. It likely provides the necessary framework or evidence to understand how generative AI impacts various job roles heterogeneously. No abstract available.	Edward W. Felten, Manav Raj, Robert Seamans
10	2023	The Short-Term Effects of Generative Artificial Intelligence on Employment: Evidence from an Online Labor Market ↗ This paper directly addresses the core question of whether generative AI substitutes for or augments labor by providing empirical evidence from an online labor market. It specifically examines the causal effects of AI exposure on employment and earnings, identifying winners and losers among workers with varying skill levels and experience. Generative artificial intelligence (AI) holds the potential to either complement workers by enhancing their productivity or substitute them. We examine the short-term effects of the recently released generative AI models (ChatGPT, DALL-E 2, and Midjourney) on the employment outcomes of freelancers on a large online platform. We find that freelancers in highly affected occupations suffer from the introduction of generative AI, experiencing reductions in both employment and earnings. We find similar effects studying the release of other image-based generative AI models. Exploring the heterogeneity by freelancers’ employment history, we do not find evidence that high-quality service, measured...	Xiang Hui, Oren Reshef, Luofeng Zhou
10	2024	The Effects of Generative AI on High Skilled Work: Evidence from Three Field Experiments with Software Developers ↗ [Title only] This paper directly addresses the project's core interest in generative AI productivity experiments by providing field experiment evidence from high-skilled workers, specifically software developers. It allows for an analysis of whether AI augments or substitutes labor and impacts wage and employment dynamics among skilled professionals, which is central to understanding distributional effects and task reorganization. No abstract available.	Zheyuan Cui, Mert Demirer, Sonia Jaffe, et al.
10	2024	Who is AI Replacing? The Impact of Generative AI on Online Freelancing Platforms ↗ This paper directly addresses the project's core questions by empirically quantifying how generative AI substitutes for labor in online freelancing, specifically impacting writing, coding, and image creation tasks. It provides crucial evidence on the distributional effects of AI by highlighting the reduction in job posts for automation-prone roles and the resulting changes in competition and wage structures for remaining workers. This paper studies the impact of generative artificial intelligence (AI) technologies on the demand for online freelancers using a large data set from a leading global freelancing platform. We identify the types of jobs that are more affected by generative AI and quantify the magnitude of the heterogeneous impact. Our findings indicate a 21% decrease in the number of job posts for automation-prone jobs related to writing and coding compared with jobs requiring manual-intensive skills within eight months after the introduction of ChatGPT. We show that the reduction in the number of job posts increases competition among freelancers, whereas the remaining automation-prone jobs are of greater...	Ozge Demirci, Jonas Hannane, Xinrong Zhu
10	2025	The Labor Market Effects of Generative Artificial Intelligence ↗ This paper directly addresses the project's core question regarding whether generative AI substitutes for labor, providing empirical evidence of reduced employment and earnings for freelancers in affected occupations. It specifically investigates the distributional effects on high-skill workers and the impact on online labor markets, which are central themes of the research. Generative artificial intelligence (AI) holds the potential to either complement workers by enhancing their productivity or substitute them. We examine the short-term effects of the recently released generative AI models (ChatGPT, DALL-E 2, and Midjourney) on the employment outcomes of freelancers on a large online platform. We find that freelancers in highly affected occupations suffer from the introduction of generative AI, experiencing reductions in both employment and earnings. We find similar effects studying the release of other image-based generative AI models. Exploring the heterogeneity by freelancers’ employment history, we do not find evidence that high-quality service, measured...	Jonathan Hartley, Filip Jolevski, Victor Hugo Cunha de Melo, et al.
10	2024	The Simple Macroeconomics of AI ↗ This paper directly addresses the project's core themes by providing a task-based macroeconomic model of AI's impact on productivity, wages, and inequality. It explicitly evaluates the aggregate labor market effects of AI, offering theoretical and empirical insights into how AI exposure translates into economywide outcomes, which is central to the researcher's investigation. This paper evaluates claims about large macroeconomic implications of new advances in AI. It starts from a task-based model of AI’s effects, working through automation and task complementarities. So long as AI’s microeconomic effects are driven by cost savings/productivity improvements at the task level, its macroeconomic consequences will be given by a version of Hulten’s theorem: GDP and aggregate productivity gains can be estimated by what fraction of tasks are impacted and average task-level cost savings. Using existing estimates on exposure to AI and productivity improvements at the task level, these macroeconomic effects appear nontrivial but modest—no more than a 0.66% increase in...	Daron Acemoğlu
10	2025	Generative AI as Seniority-Biased Technological Change: Evidence from U.S. Résumé and Job Posting Data ↗ This paper directly addresses the project's core question regarding the impact of AI on entry-level workers and traditional career ladders, providing causal evidence of seniority-biased technological change. It empirically links GenAI adoption to reduced demand for junior labor through task displacement, offering key insights into distributional effects and firm-level labor reorganization. We study whether generative AI (GenAI) constitutes seniority-biased technological change, disproportionately reducing demand for junior workers. We develop a conceptual framework in which GenAI adoption reduces junior labor demand through task displacement and labor-saving productivity gains. We test the framework's mechanisms and implications using U.S. résumé data covering 65 million workers at more than 280,000 firms (2015--2025), allowing us to track firm-level employment by seniority. GenAI adoption is identified through text analysis that detects ‘GenAI integrator’ job postings, signaling active GenAI implementation by firms. Following adoption, junior employment declines sharply in...	Guy Lichtinger, Seyed Mahdi Hosseini Maasoum
10	2025	Augmenting or Automating Labor? The Effect of AI Development on New Work, Employment, and Wages ↗ This paper directly addresses the core questions regarding whether AI augments or substitutes for labor and its distinct effects on employment and wages by skill level. It provides empirical evidence on the distributional consequences of AI, specifically highlighting how automation and augmentation technologies differently impact low-skilled versus high-skilled occupations, which is central to the project's investigation of winners, losers, and inequality. Artificial intelligence (AI) is reshaping the labor market by changing the task content of occupations. This study investigates the impact of AI development on the emergence of new work, employment, and wages in the United States from 2015 to 2022. I develop innovative methods to measure occupational and industry exposure to AI technologies that substitute labor (automation AI ) or enhance workers' output (augmentation AI), and to identify new work (i.e., new job titles). To address endogeneity, I use instrumental variable estimators, leveraging AI development in countries with limited economic ties to the United States. The findings indicate that automation AI negatively impacts new work...	David Marguerit
10	2023	Automation: Theory, Evidence, and Outlook ↗ This paper is highly relevant as it directly addresses the project's core themes by reviewing the task-based framework and its implications for labor markets, wages, and productivity. It provides essential theoretical and empirical context regarding automation's substitution effects and displacement mechanisms, which are central to understanding AI's impact on workers. This article reviews the literature on automation and its impact on labor markets, wages, factor shares, and productivity. I first introduce the task model and explain why this framework offers a compelling way to think about recent labor market trends and the effects of automation technologies. The task model clarifies that automation technologies operate by substituting capital for labor in a widening range of tasks. This substitution reduces costs, creating a positive productivity effect, but it also reduces employment opportunities for workers displaced from automated tasks, creating a negative displacement effect. I survey the empirical literature and conclude that there is wide...	Pascual Restrepo
10	2025	The Labor Market Effects of Generative AI: A Difference-in-Differences Analysis of AI Exposure ↗ This paper directly addresses the project's core questions regarding the causal effects of AI on worker productivity, employment, and aggregate labor market outcomes using a rigorous difference-in-differences design. It provides key empirical evidence on how AI exposure impacts output and labor share, while distinguishing between augmentative and substitutive effects based on collaboration requirements. Does artificial intelligence (AI) increase productivity---and does it displace workers? We examine aggregate effects using administrative data covering essentially all U.S. employers in a difference-in-differences design exploiting occupational AI exposure across industries and states. A one standard deviation increase in exposure raises output by 7%, with effects emerging in 2021 when enterprise AI tools entered the market. Employment effects follow the same timing but diverge by exposure type: where AI likely requires human collaboration, employment rises 4%; where AI can perform tasks independently, we find no significant employment effect. Results are robust to state-by-year and...	Andrew Johnston, Christos Makridis
10	2026	Who is using AI to code? Global diffusion and impact of generative AI ↗ This paper directly addresses the project's core questions regarding AI's causal effects on worker productivity and its distributional impact by skill level and career stage. It provides critical empirical evidence on how generative AI augments labor for experienced workers while potentially harming entry-level workers, thereby reshaping career ladders and inequality in the labor market. Generative coding tools promise big productivity gains, but uneven uptake could widen skill and income gaps. We train a neural classifier to spot artificial intelligence (AI)-generated Python functions in more than 30 million GitHub commits by 160,097 software developers, tracking how fast, and where, these tools take hold. Currently, AI writes an estimated 29% of Python functions in the US-a shrinking lead over other countries. We estimate that quarterly output, measured in online code contributions, consequently increased by 3.6%. AI seems to benefit experienced, senior-level developers: They increased productivity and more readily expanded into new domains of software development. By...	Simone Daniotti, Johannes Wachs, Xiangnan Feng, et al.
10	2025	Generative AI and Firm Productivity: Field Experiments in Online Retail ↗ This paper directly addresses the project's core question on the causal effects of generative AI on worker and firm productivity through rigorous field experiments. It provides high-quality empirical evidence on how AI tools augment labor in specific tasks, such as customer service and product matching, aligning perfectly with the project's focus on productivity impacts and task-based frameworks. We quantify the short-term impact of Generative Artificial Intelligence (GenAI) on sales performance through a series of large-scale randomized field experiments involving millions of users and products at a leading cross-border online retail platform. Over 2023-2024, the platform integrated GenAI into seven consumer-facing business workflows spanning customer service, consumer-product matching, advertising, and seller services. We find that GenAI adoption increases sales in most workflows, with effects ranging from no detectable impact to $16.3\%$, depending on GenAI's marginal contribution relative to baseline firm practices. Across the four GenAI applications with positive sales effects...	Lu Fang, Zhe Yuan, Kaifu Zhang, et al.
10	2024	Towards the Terminator Economy: Assessing Job Exposure to AI through LLMs ↗ This paper directly addresses the project's core questions by developing and validating new measures of AI exposure and replacement at the task level using large language models. It provides critical empirical evidence on how AI exposure varies by skill and occupation, and analyzes the causal relationship between AI adoption and labor market outcomes like wages and employment. AI and related technologies are reshaping jobs and tasks, either by automating or augmenting human skills in the workplace. Many researchers have been working on estimating if and to what extent jobs and tasks are exposed to the risk of being automatized by AI-related technologies. Our work tackles this issue through a data-driven approach by: (i) developing a reproducible framework that uses cutting-edge open-source large language models to assess the current capabilities of AI and robotics in performing job-related tasks; (ii) formalizing and computing a measure of AI exposure by occupation, the Task Exposure to AI (TEAI) index, and a measure of Task Replacement by AI (TRAI), both...	Emilio Colombo, Fabio Mercorio, Mario Mezzanzanica, et al.
10	2025	Generative AI as Routine-Biased Technical Change? Evidence from a Field Experiment in Central Banking ↗ This paper directly addresses the project's core question of whether AI tools augment or substitute for labor by providing rigorous experimental evidence on task-based complementarities in a real-world setting. It specifically investigates how AI exposure varies by task type and skill level, and analyzes how firms reorganize work to maximize productivity, aligning perfectly with the project's themes of AI skill complementarity and task reorganization. We examine how generative AI impacts productivity across the task-based framework using a field experiment at the National Bank of Slovakia. In our experiment, we randomly assign generative AI access to central bank employees completing workplace tasks that mirror the theoretical task-based framework. Our results indicate that generative AI access leads to large improvements in both quality and efficiency for the majority of participants. We find a strong complementarity between generative AI and non-routine work, both on average and for most participants. We also find some support for generative AI as both cognitive-biased and specialist-biased, though smaller in magnitude than our tests...	Ales Marsal, Patryk Perkowski
10	2024	AI Adoption and Firm Demand for Workers and Skills: Insights from Online Job Postings ↗ This paper directly addresses the project's core question of whether AI tools augment or substitute for labor by analyzing hiring trends in AI-adopting firms. It provides empirical evidence on how AI adoption affects labor demand, skill requirements, and the composition of jobs, aligning perfectly with the project's focus on distributional effects and task reorganization. The latest Artificial Intelligence (AI) tools can perform some of the complex tasks that highly skilled and well-paid workers perform. To investigate their effects on demand for workers and skills, we compared hiring trends in Australian firms that were adopting AI and those that were not. Job postings grew significantly faster in firms that had adopted AI, even after controlling for firm size, geography and industry. This accelerated growth in job postings included occupations that were highly exposed to AI. The number of skills sought in job postings was also growing faster for AI exposed occupations, especially if they were being recruited by AI adopting firms. Some formerly non-AI...	Claire Mason, Haohui Chen, David Evans
10	2025	Tasks at work: comparative advantage, technology and labor demand ↗ This paper directly addresses the project's core task-based framework by explaining how technology, including AI, reallocates tasks between labor and capital through substitution and creation mechanisms. It provides the essential theoretical foundation for analyzing whether AI augments or substitutes for labor and how this reshapes labor demand, productivity, and inequality. This chapter reviews recent advances in the task model and shows how this framework can be put to work to understand trends in the labor market in recent decades. Production in each industry requires the completion of various tasks that can be assigned to workers with different skills or to capital. Factors of production have well-defined comparative advantage across tasks, which governs substitution patterns. Technological change can: (1) augment a specific labor type—e.g., increase the productivity of labor in tasks it is already performing; (2) augment capital; (3) automate work by enabling capital to perform tasks previously allocated to labor; (4) create new tasks. The task model...	Daron Acemoglu, Fredric Kong, Pascual Restrepo
10	2025	Generative AI in Action: Field Experimental Evidence on Worker Performance in E-Commerce Customer Service Operations ↗ This paper directly addresses the project's core questions by providing causal evidence on how generative AI augments labor through a large-scale field experiment on worker productivity. It explicitly examines distributional effects by showing how AI narrows performance gaps for low-skilled workers while potentially harming top performers, thereby illuminating the dynamics of AI skill complementarity and task reorganization. In collaboration with Alibaba, this study leverages a large-scale field experiment to assess the impact of a generative AI assistant on worker performance in e-commerce after-sales service. Human agents providing digital chat support were randomly assigned with access to a gen AI assistant that offered two core functions: diagnosis of customer issues and solution proposals, presented as text messages. Agents retained discretion to adopt, modify, or disregard AI-generated messages. To evaluate gen AI's impact, we estimate both the intention-to-treat (ITT) effect of gen AI access and the local average treatment effect (LATE) of gen AI usage. Results show that gen AI significantly improved...	Ni Xiao, Yiwei Wang, Tianjun Feng, et al.
10	2025	AI and jobs. A review of theory, estimates, and evidence ↗ This review directly addresses the project's core questions by synthesizing theory and empirical evidence on AI's impact on employment, productivity, and inequality. It specifically examines task-based frameworks, the differential effects on novice versus skilled workers, and the mechanisms of augmentation versus substitution across various labor market contexts. Generative AI is altering work processes, task composition, and organizational design, yet its effects on employment and the macroeconomy remain unresolved. In this review, we synthesize theory and empirical evidence at three levels. First, we trace the evolution from aggregate production frameworks to task- and expertise-based models. Second, we quantitatively review and compare (ex-ante) AI exposure measures of occupations from multiple studies and find convergence towards high-wage jobs. Third, we assemble ex-post evidence of AI's impact on employment from randomized controlled trials (RCTs), field experiments, and digital trace data (e.g., online labor platforms, software repositories)...	R. Maria del Rio-Chanona, Ekkehard Ernst, Rossana Merola, et al.
10	2025	Towards the Terminator Economy: Assessing Job Exposure to Ai Through Llms ↗ This paper directly addresses the project's core theme of measuring AI exposure across occupations by developing and validating the TEAI and TRAI indices using large language models. It provides critical empirical evidence on the causal link between AI exposure and labor market outcomes like employment and wages, while also analyzing whether AI acts as a substitute or complement to labor in high-skill jobs. AI and related technologies are reshaping jobs and tasks, either by automating or augmenting human skills in the workplace. Many researchers have been working on estimating if and to what extent jobs and tasks are exposed to the risk of being automatized by AI-related technologies. Our work tackles this issue through a data-driven approach by: (i) developing a reproducible framework that uses cutting-edge open-source large language models to assess the current capabilities of AI and robotics in performing job-related tasks; (ii) formalizing and computing a measure of AI exposure by occupation, the Task Exposure to AI (TEAI) index, and a measure of Task Replacement by AI (TRAI) index, both...	Emilio Colombo, Fabio Mercorio, Mario Mezzanzanica, et al.
10	2024	Review of: "AI Adoption and Firm Demand for Workers and Skills: Insights from Online Job Postings" ↗ [Title only] This paper directly addresses the core theme of firm AI adoption and its impact on labor demand using online job postings, which is a primary data source for measuring AI exposure and skill complementarity. It likely provides critical insights into how firms are reorganizing work and shifting skill requirements in response to AI, aligning perfectly with the project's focus on distributional effects and task reorganization. No abstract available.	Marina Mendes Tavares
10	2026	Generative AI and labour productivity: A quasi experiment on coding ↗ This paper directly addresses the core project question by providing causal evidence on how generative AI affects worker productivity, specifically finding significant gains for entry-level developers while noting limited impact on senior staff. It aligns perfectly with the project's focus on task-based frameworks, augmentation versus substitution, and the differential effects of AI on workers at different skill levels. This paper examines the impact of generative artificial intelligence (Gen AI) on labour productivity through a quasi-experiment involving software developers. In September 2023, Ant Group introduced CodeFuse, a large language model (LLM) designed to support coding tasks. While some programmer teams began using CodeFuse, others were not made aware of its release. Exploiting this natural variation in exposure, we identify comparable treatment and control groups of programmers to estimate the causal effect of Gen AI adoption on productivity. We find that the use of CodeFuse increases code output by over 50%. However, these productivity gains are statistically significant only among junior or...	Leonardo Gambacorta, Han Qiu, Shuo Shan, et al.
10	2026	Creation, validation, obsolescence: observed evidence of AI-driven labor market displacement, 2020–2025 ↗ This paper directly addresses the project's core questions by providing empirical evidence on AI-driven labor market displacement, wage effects, and distributional impacts across occupations and skill levels since 2020. It specifically analyzes the contraction of entry-level and mid-level roles, the emergence of an AI-augmentation wage premium, and the broader implications for inequality and task reorganization. Background The successive releases of GPT-3 (May 2020) and ChatGPT (November 2022) have been widely hypothesized to constitute inflection points in the automation of cognitive labor. Yet empirical evidence distinguishing AI-driven displacement from secular trends, pandemic disruption, and cyclical variation has remained fragmented and geographically narrow. Methods Following PRISMA 2020 guidelines, we systematically searched six academic databases (Scopus, Web of Science, EconLit, SSRN, IEEE Xplore, Google Scholar) for empirical studies documenting observed—not predicted—labor market changes since 2020. From 1,847 initial records, 94 studies meeting inclusion criteria were retained for...	Nassim Dehouche
10	2026	Robots Reinstate, AI Doesn't: Asymmetric Task Creation Across Automation Frontiers * ↗ This paper directly addresses the core question of whether AI augments or substitutes for labor by distinguishing its effects from physical automation using a task-based framework. It provides crucial empirical evidence on the asymmetric creation of complementary tasks, highlighting how AI's non-spatial nature leads to different wage trajectories and employment outcomes compared to robots. Robot-exposed occupations follow V-shaped wage trajectories-decline then recoverywhile AI-exposed occupations follow L-shapes with no recovery. I develop a twofrontier task model in which the direction of automation and the rate of complementary task creation are jointly determined. Physical automation produces spatially bundled complements-maintenance, monitoring, quality control-because automated and human tasks are co-located , cognitive automation produces fewer complements because its outputs are digital and non-spatial. Calibrating the model to structurally estimated reinstatement rates identifies a complementarity ratio of 4.88: each unit of robot frontier advancement generates...	Carlos Chavez
9	2016	The future of employment: How susceptible are jobs to computerisation? ↗ This foundational paper directly addresses the project's core theme of measuring AI exposure by introducing a seminal framework for estimating the probability of job computerisation across occupations. It provides essential background on how susceptibility to automation correlates with wages and education, which is critical for understanding the distributional effects and task-based analysis central to the researcher's inquiry. We examine how susceptible jobs are to computerisation. To assess this, we begin by implementing a novel methodology to estimate the probability of computerisation for 702 detailed occupations, using a Gaussian process classifier. Based on these estimates, we examine expected impacts of future computerisation on US labour market outcomes, with the primary objective of analysing the number of jobs at risk and the relationship between an occupations probability of computerisation, wages and educational attainment.	Carl Benedikt Frey, Michael A. Osborne
9	2022	Artificial Intelligence and Jobs: Evidence from Online Vacancies ↗ This paper directly addresses the project's core questions regarding AI exposure measurement and the causal effects of AI on labor markets by analyzing online vacancy data. It provides critical empirical evidence on how AI adoption reshapes hiring patterns, skill requirements, and the balance between substitution and augmentation in specific occupations. We study the impact of artificial intelligence (AI) on labor markets using establishment-level data on the near universe of online vacancies in the United States from 2010 onward. There is rapid growth in AI-related vacancies over 2010–18 that is driven by establishments whose workers engage in tasks compatible with AI’s current capabilities. As these AI-exposed establishments adopt AI, they simultaneously reduce hiring in non-AI positions and change the skill requirements of remaining postings. While visible at the establishment level, the aggregate impacts of AI-labor substitution on employment and wage growth in more exposed occupations and industries is currently too small to be...	Daron Acemoğlu, David Autor, Jonathon Hazell, et al.
9	2019	Artificial Intelligence, Automation, and Work ↗ This paper directly addresses the core task-based framework and causal mechanisms (displacement vs. productivity effects) central to the project's investigation of AI's impact on labor markets. It provides the theoretical foundation for understanding how AI reshapes employment, wages, and inequality through task reorganization and skill complementarity. We summarize a framework for the study of the implications of automation and AI on the demand for labor, wages, and employment.Our task-based framework emphasizes the displacement effect that automation creates as machines and AI replace labor in tasks that it used to perform.This displacement effect tends to reduce the demand for labor and wages.But it is counteracted by a productivity effect, resulting from the cost savings generated by automation, which increase the demand for labor in non-automated tasks.The productivity effect is complemented by additional capital accumulation and the deepening of automation (improvements of existing machinery), both of which further increase the...	Daron Acemoğlu, Pascual Restrepo
9	2021	Occupational, industry, and geographic exposure to artificial intelligence: A novel dataset and its potential uses ↗ This paper directly addresses the project's core theme of measuring AI exposure by introducing a novel dataset (AIOE, AIIE, AIGE) for occupations, industries, and geographies. It provides the essential measurement infrastructure needed to empirically analyze how AI reshapes labor markets and affects workers across different sectors and locations. Abstract Research Summary We create and validate a new measure of an occupation's exposure to AI that we call the AI Occupational Exposure (AIOE). We use the AIOE to construct a measure of AI exposure at the industry level, which we call the AI Industry Exposure (AIIE) and a measure of AI exposure at the county level, which we call the AI Geographic Exposure (AIGE). We also describe several ways in which the AIOE can be used to create firm level measures of AI exposure. We validate the measures and describe how they can be used in different applications by management, organization and strategy scholars. Managerial Summary Although artificial intelligence (AI) promises to spur economic...	Edward W. Felten, Manav Raj, Robert Seamans
9	2020	Artificial Intelligence as Augmenting Automation: Implications for Employment ↗ This paper directly addresses the core question of whether AI tools augment or substitute for labor, a central theme of the project. It critically examines the implications for employment, aligning closely with the research focus on causal effects and distributional outcomes. There has been great concern in recent years that artificial intelligence (AI) may cause widespread unemployment, but proponents say that AI augments existing jobs. Both of these positions have sub...	Feichin Tschang, Esteve Almirall
9	2013	The “task approach” to labor markets: an overview ↗ This paper provides a foundational overview of the task-based framework, which is a core theme of the research project for understanding how technology reshapes labor markets. It directly addresses the measurement and conceptualization of AI exposure by analyzing the allocation of tasks between labor and capital, a mechanism central to determining whether AI acts as a substitute or complement. An emerging literature argues that changes in the allocation of workplace “tasks” between capital and labor, and between domestic and foreign workers, has altered the structure of labor demand in industrialized countries and fostered employment polarization—that is, rising employment in the highest and lowest paid occupations. Analyzing this phenomenon within the canonical production function framework is challenging, however, because the assignment of tasks to labor and capital in the canonical model is essentially static. This essay sketches an alternative model of the assignment of skills to tasks based upon comparative advantage, reviews key conceptual and practical challenges that...	David Autor
9	2021	The impact of artificial intelligence on labor productivity ↗ This paper directly addresses the project's core question regarding the causal effects of AI on worker productivity by analyzing the impact of AI patenting on firm-level labor productivity. It provides empirical evidence on how AI adoption influences output, specifically highlighting differential effects across firm size and sectors, which is central to understanding the economic impact of AI. Abstract Recent evidence indicates an upsurge in artificial intelligence and robotics (AI) patenting activities in the latest years, suggesting that solutions based on AI technologies might have started to exert an effect on the economy. We test this hypothesis using a worldwide sample of 5257 companies having filed at least a patent related to the field of AI between 2000 and 2016. Our analysis shows that, once controlling for other patenting activities, AI patent applications generate an extra-positive effect on companies’ labor productivity. The effect concentrates on SMEs and services industries, suggesting that the ability to quickly readjust and introduce AI-based applications in the...	Giacomo Damioli, Vincent Van Roy, Dániel Vértesy
9	2023	The Impact of AI on Developer Productivity: Evidence from GitHub Copilot ↗ This paper directly addresses the core project theme of generative AI productivity experiments by providing controlled experimental evidence on how AI tools impact worker productivity in a specific high-skill occupation. It also touches on distributional effects and the future of work by suggesting AI may facilitate transitions into software development, relevant to the project's interest in entry-level workers and career ladders. Generative AI tools hold promise to increase human productivity. This paper presents results from a controlled experiment with GitHub Copilot, an AI pair programmer. Recruited software developers were asked to implement an HTTP server in JavaScript as quickly as possible. The treatment group, with access to the AI pair programmer, completed the task 55.8% faster than the control group. Observed heterogenous effects show promise for AI pair programmers to help people transition into software development careers.	Sida Peng, Eirini Kalliamvakou, Peter Cihon, et al.
9	2021	Artificial intelligence and employment ↗ This paper directly addresses the project's core question regarding the causal effects of AI exposure on employment by analyzing cross-country data and occupation-level AI indicators. It provides key empirical evidence on whether AI augments or substitutes labor, highlighting the critical role of digital skills in determining outcomes, which aligns perfectly with the themes of labor market reshaping and distributional effects. Recent years have seen impressive advances in artificial intelligence (AI) and this has stoked renewed concern about the impact of technological progress on the labour market, including on worker displacement. This paper looks at the possible links between AI and employment in a cross-country context. It adapts the AI occupational impact measure developed by Felten, Raj and Seamans (2018[1]; 2019[2]) – an indicator measuring the degree to which occupations rely on abilities in which AI has made the most progress – and extends it to 23 OECD countries. The indicator, which allows for variations in AI exposure across occupations, as well as within occupations and across countries, is then...	Alexandre Georgieff, Raphaela Hyee
9	2023	How will Language Modelers like ChatGPT Affect Occupations and Industries? ↗ [Title only] This title directly addresses the core project question of how generative AI, specifically large language models, impacts labor markets across different occupations and industries. It aligns perfectly with the themes of AI exposure measurement, task-based frameworks, and distributional effects by explicitly linking the technology to specific economic sectors and job types. No abstract available.	Edward W. Felten, Manav Raj, Robert Seamans
9	2021	Does artificial intelligence affect the pattern of skill demand? Evidence from Chinese manufacturing firms ↗ This paper directly addresses the project's core question regarding who the winners and losers are by examining how AI adoption shifts skill demand towards high-skilled labor and away from low-skilled workers. It provides causal evidence on the substitution vs. complementarity debate using a quasi-natural experiment, which aligns with the project's focus on distributional effects and labor market impacts. Abstract In view of the recent penetration of artificial intelligence (AI) into production activities, we undertake a quasi-natural experiment to identify its impact on employment at different skill levels using micro-enterprise data from Chinese manufacturing during 2011–2017. Employing a robust difference-in-differences method with propensity score matching, we investigate the heterogeneous impact of AI adoption upon different skills across three dimensions — geographical regions, enterprise types, and the length of time since the adoption of AI. We find that AI reduces the relative demand for low-skilled labor across all regions in China, while increasing the relative demand for...	Mengmeng Xie, Lin Ding, Yan Xia, et al.
9	2024	The Short-Term Effects of Generative Artificial Intelligence on Employment: Evidence from an Online Labor Market ↗ This paper directly addresses the project's core question regarding whether AI tools augment or substitute for labor by providing empirical evidence of negative employment and earnings effects for freelancers in affected occupations. It offers critical insights into the distributional effects of generative AI, specifically highlighting that even high-skill workers may be disproportionately impacted, which relates to the project's interest in winners, losers, and task reorganization. Generative artificial intelligence (AI) holds the potential to either complement workers by enhancing their productivity or substitute them. We examine the short-term effects of the recently released generative AI models (ChatGPT, DALL-E 2, and Midjourney) on the employment outcomes of freelancers on a large online platform. We find that freelancers in highly affected occupations suffer from the introduction of generative AI, experiencing reductions in both employment and earnings. We find similar effects studying the release of other image-based generative AI models. Exploring the heterogeneity by freelancers’ employment history, we do not find evidence that high-quality service, measured...	Xiang Hui, Oren Reshef, Luofeng Zhou
9	2023	Technological Change and the Consequences of Job Loss ↗ This paper directly addresses the project's core interest in the distributional effects of technological change on wages and employment, specifically focusing on the persistent earnings losses experienced by displaced workers. By linking skill requirements to technological change and occupational mobility, it provides critical empirical evidence on how technology reshapes labor market outcomes and inequality, a key theme in the researcher's project. We examine the role of technological change in explaining the large and persistent decline in earnings following job loss. Using detailed skill requirements from the near universe of online vacancies, we estimate technological change by occupation and find that technological change accounts for 45 percent of the decline in earnings after job loss. Technological change lowers earnings after job loss by requiring workers to have new skills to perform newly created jobs in their prior occupation. When workers lack the required skills, they move to occupations where their skills are still employable but are paid a lower wage. (JEL J24, J31, J63, O33)	J. Carter Braxton, Bledi Taska
9	2022	Artificial Intelligence and Employment: New Cross-Country Evidence ↗ This paper directly addresses the core question of causal effects of AI on employment by providing cross-country evidence using an established AI exposure measure. It specifically investigates the distributional effects by skill level and computer use, examining whether AI augments or substitutes labor and how it impacts different worker groups. Recent years have seen impressive advances in artificial intelligence (AI) and this has stoked renewed concern about the impact of technological progress on the labor market, including on worker displacement. This paper looks at the possible links between AI and employment in a cross-country context. It adapts the AI occupational impact measure developed by Felten, Raj and Seamans—an indicator measuring the degree to which occupations rely on abilities in which AI has made the most progress—and extends it to 23 OECD countries. Overall, there appears to be no clear relationship between AI exposure and employment growth. However, in occupations where computer use is high, greater exposure to...	Alexandre Georgieff, Raphaela Hyee
9	2019	Automation and New Tasks: How Technology Displaces and Reinstates Labor ↗ This paper is highly relevant as it establishes the foundational task-based framework used to analyze how technology displaces and reinstates labor, which is a core theme of the project. It directly addresses the mechanisms of AI's impact on employment and wages through the lenses of substitution and complementarity within a structural economic model. We present a framework for understanding the effects of automation and other types of technological changes on labor demand, and use it to interpret changes in US employment over the recent past. At the center of our framework is the allocation of tasks to capital and labor the task content of production. Automation, which enables capital to replace labor in tasks it was previously engaged in, shifts the task content of production against labor because of a displacement effect. As a result, automation always reduces the labor share in value added and may reduce labor demand even as it raises productivity.The effects of automation are counterbalanced by the creation of new tasks in which...	Daron Acemoğlu, Pascual Restrepo
9	2017	Wage Shocks and the Technological Substitution of Low‐wage Jobs ↗ This paper directly addresses the project's core interest in how technological changes substitute for specific labor tasks, specifically focusing on low-wage cognitive roles. It provides crucial empirical evidence on the causal mechanisms of task-based substitution and its distributional effects on wages and employment, which are central to understanding AI's impact on labor markets. We extend the task-based empirical framework used in the job polarization literature to analyze the susceptibility of low-wage employment to technological substitution. We find that increases in the cost of low-wage labor, via minimum wage hikes, lead to relative employment declines at cognitively routine occupations but not manually-routine or non-routine low-wage occupations. This suggests that low-wage routine cognitive tasks are susceptible to technological substitution. While the short-run employment consequence of this reshuffling on individual workers is economically small, due to concurrent employment growth in other low-wage jobs, workers previously employed in cognitively routine...	Daniel Aaronson, Brian J. Phelan
9	2020	Artificial Intelligence and High-Skilled Work: Evidence from Analysts ↗ This paper directly addresses the project's core question of how AI affects high-skilled workers by examining task reallocation, specifically the shift toward soft skills and coverage changes. It provides crucial evidence on whether AI acts as a substitute or complement to labor and details the distributional consequences for worker wages and career paths. Policymakers fear artificial intelligence (AI) will disrupt labor markets, especially for high-skilled workers. We investigate this concern using novel, task-specific data for security analysts. Exploiting variation in AI's power across stocks, we show analysts with portfolios that are more exposed to AI are more likely to reallocate efforts to soft skills, shift coverage towards low AI stocks, and even leave the profession. Analyst departures disproportionately occur among highly accurate analysts, leaving for non-research jobs. Reallocating efforts toward tasks that rely on social skills improve consensus forecasts. However, increased exposure to AI reduces the novelty in analysts'...	Jillian Grennan, Roni Michaely
9	2024	The unequal adoption of ChatGPT exacerbates existing inequalities among workers ↗ This paper directly addresses the core themes of AI adoption disparities and distributional effects by examining how ChatGPT usage varies across gender, age, and experience levels. It provides empirical evidence on how unequal access and adoption patterns may exacerbate existing inequalities among workers, which is central to the project's inquiry into who the winners and losers of AI are. We study the adoption of ChatGPT, the icon of Generative AI, using a large-scale survey linked to comprehensive register data in Denmark. Surveying 18,000 workers from 11 exposed occupations, we document that ChatGPT is widespread, especially among younger and less-experienced workers. However, substantial inequalities have emerged. Women are 16 percentage points less likely to have used the tool for work. Furthermore, despite its potential to lift workers with less expertise, users of ChatGPT earned slightly more already before its arrival, even given their lower tenure. Workers see a substantial productivity potential in ChatGPT but are often hindered by employer restrictions and a...	Anders Humlum, Emilie Vestergaard
9	2022	What skills and abilities can automation technologies replicate and what does it mean for workers? ↗ This paper directly addresses the project's core theme of measuring AI exposure through a granular, task-based framework that links automatable skills to specific occupations. It provides critical insights into how AI reshapes labor markets by identifying high-level cognitive tasks as automatable and explaining how work organization changes rather than jobs simply disappearing. This paper exploits novel data on the degree of automatability of approximately 100 skills and abilities collected through an original survey of experts in AI, and link them to occupations using information on skill and ability requirements extracted from O*NET. Similar to previous studies, this allows gauging the number of jobs potentially affected by automation and the workers who are most at risk of automation. The focus on the automatability of skills and abilities as opposed to entire occupations permits a direct assessment of the share of highly automatable and bottleneck tasks in each occupation. The study finds that thanks to advances in AI and robotics, several high-level cognitive...	Julie Lassébie, Glenda Quintini
9	2019	The Effect of Artificial Intelligence on Human Labor: An Ability-Based Approach ↗ This paper directly addresses the project's core inquiry by constructing an AI exposure measure and analyzing its causal effects on employment, wages, and inequality across different skill levels. It explicitly investigates whether AI acts as a substitute or complement to labor, which is a central mechanism in the researcher's framework. While artificial intelligence (AI) promises to spur economic growth, there is concern that this may come at the expense of human labor. We utilize data on advances in AI together with occupational definitions to construct an occupation-level measure of the impact of AI, and use this measure to investigate whether and under what circumstances AI may act as a substitute or a complement to labor. We provide broad evidence that occupations impacted by AI may see a decline in wages, but growth in employment, and that this is particularly the case for occupations with complementary skills & technologies. In addition, high-income occupations experience a growth in employment, suggesting that AI...	Edward W. Felten, Manav Raj, Robert Seamans
9	2024	How Does Artificial Intelligence Shape Audit Firms? ↗ This paper directly addresses the core questions of whether AI substitutes for or augments labor, providing causal evidence on employment effects and skill demand changes in a specific profession. It also examines firm reorganization and the impact on different worker levels, aligning closely with the project's focus on distributional effects and task-based frameworks. Does artificial intelligence (AI) displace auditors? We exploit the staggered hiring of AI employees at audit office locations across the United States as a proxy for the use of AI at local audit offices. The main findings are as follows. First, relative to audit offices that do not yet hire AI employees, those that do hire AI employees have a 4.3% increase in the number of auditor jobs, particularly among junior and midlevel auditors. Second, using AI is associated with an increased demand for soft skills (e.g., cognitive skills) in auditor jobs. Third, audit offices that use AI have more accurate going concern and internal control opinions. Semistructured interviews of 11 seasoned audit...	Kelvin Law, M.D. Shen
9	2025	Who Is AI Replacing? The Impact of Generative AI on Online Freelancing Platforms ↗ This paper directly addresses the project's core questions by analyzing the causal effects of generative AI on employment and wages within online labor markets, a key context specified in the research themes. It provides empirical evidence on which occupations and tasks are substituted by AI, detailing the heterogeneous impact on job posts, pay, and worker competition. This paper studies the impact of generative artificial intelligence (AI) technologies on the demand for online freelancers using a large data set from a leading global freelancing platform. We identify the types of jobs that are more affected by generative AI and quantify the magnitude of the heterogeneous impact. Our findings indicate a 21% decrease in the number of job posts for automation-prone jobs related to writing and coding compared with jobs requiring manual-intensive skills within eight months after the introduction of ChatGPT. We show that the reduction in the number of job posts increases competition among freelancers, whereas the remaining automation-prone jobs are of greater...	Ozge Demirci, Jonas Hannane, Xinrong Zhu
9	2024	Automation: Theory, Evidence, and Outlook ↗ This paper directly addresses the core task-based framework and theoretical mechanisms of labor market impact central to the project, specifically examining automation as a substitute for labor. It provides a comprehensive review of the displacement and productivity effects that are fundamental to understanding how AI reshapes employment and wages. This article reviews the literature on automation and its impact on labor markets, wages, factor shares, and productivity. I first introduce the task model and explain why this framework offers a compelling way to think about recent labor market trends and the effects of automation technologies. The task model clarifies that automation technologies operate by substituting capital for labor in a widening range of tasks. This substitution reduces costs, creating a positive productivity effect, but it also reduces employment opportunities for workers displaced from automated tasks, creating a negative displacement effect. I survey the empirical literature and conclude that there is wide...	Pascual Restrepo
9	2025	The Cybernetic Teammate: A Field Experiment on Generative AI Reshaping Teamwork and Expertise ↗ This paper directly addresses the project's core question of whether AI augments or substitutes labor by providing rigorous field experiment evidence on how generative AI reshapes teamwork and expertise sharing. It offers critical insights into firm-level work reorganization and the changing dynamics of professional collaboration, which are central to understanding the causal effects of AI on productivity and labor market structures. We examine how artificial intelligence transforms the core pillars of collaboration-performance, expertise sharing, and social engagement-through a pre-registered field experiment with 776 professionals at Procter & Gamble, a global consumer packaged goods company.Working on real product innovation challenges, professionals were randomly assigned to work either with or without AI, and either individually or with another professional in new product development teams.Our findings reveal that AI significantly enhances performance: individuals with AI matched the performance of teams without AI, demonstrating that AI can effectively replicate certain benefits of human collaboration.Moreover, AI...	Fabrizio Dell’Acqua, Charles Ayoubi, Hila Lifshitz, et al.
9	2023	New technologies and jobs in Europe ↗ This paper directly addresses the project's core question regarding the causal effects of AI on employment shares and inequality by analyzing occupation-level exposure across European labor markets. It provides critical macro-level evidence on whether AI acts as a complement or substitute for labor, particularly regarding skilled and younger workers, which aligns with the distributional and aggregate labor market themes of the research. We examine the link between labour market developments and new technologies such as artificial intelligence (AI) and software in 16 European countries over the period 2011-2019. Using data for occupations at the 3-digit level in Europe, we find that on average employment shares have increased in occupations more exposed to AI. This is particularly the case for occupations with a relatively higher proportion of younger and skilled workers. This evidence is in line with the Skill-Biased Technological Change theory. While there is heterogeneity across countries, very few countries show a decline in the employment shares of occupations more exposed to AI-enabled automation. Country...	Stefania Albanesi, António Dias da Silva, Juan F. Jimeno, et al.
9	2024	Tracking Firm Use of AI in Real Time: A Snapshot from the Business Trends and Outlook Survey ↗ This paper directly addresses the project's core theme of firm AI adoption by providing timely, real-time empirical estimates of how widely and specifically businesses are integrating AI. It further explores key mechanisms such as task substitution, organizational reorganization, and the resulting effects on employment expansion, offering crucial evidence on the initial macroeconomic and labor market impacts of AI. Timely and accurate measurement of AI use by firms is both challenging and crucial for understanding the impacts of AI on the U.S. economy.We provide new, real-time estimates of current and expected future use of AI for business purposes based on the Business Trends and Outlook Survey for September 2023 to February 2024.During this period, bi-weekly estimates of AI use rate rose from 3.7% to 5.4%, with an expected rate of about 6.6% by early Fall 2024.The fraction of workers at businesses that use AI is higher, especially for large businesses and in the Information sector.AI use is higher in large firms but the relationship between AI use and firm size is non-monotonic.In contrast, AI use...	Kathryn R. Bonney, Cory Breaux, Cathy Buffington, et al.
9	2024	AI-driven capital-skill complementarity: Implications for skill premiums and labor mobility ↗ This paper directly addresses the project's core theme of AI skill complementarity versus substitution by proposing a modified capital-skill complementarity hypothesis. It provides a structural framework to explain how AI adoption influences labor mobility and skill premiums, which are central to understanding the distributional effects of AI on workers. To explain the newly discovered coevolution of capital and labor structures, this study presents a modified capital-skill complementarity hypothesis within the framework of structural transformation. We propose that artificial intelligence (AI) and skilled labor exhibit relative complementarity. Specifically, advancements in AI services or AI-enhanced technologies incentivize the mobilization of skilled labor across different industry sectors. The direction of this labor movement is contingent upon variations across industry sectors, including AI output elasticity and the substitutability between AI and conventional production methods. This structural transformation process also induces...	Shuo Wang, Y. G. Wang, Chengyou Li
9	2023	Generative Ai at Work ↗ This paper directly addresses the project's core question of whether AI tools augment or substitute for labor by providing empirical evidence that generative AI significantly boosts worker productivity in customer support. It also closely aligns with the theme of distributional effects by demonstrating how AI adoption differentially impacts workers based on skill level and experience, benefiting lower-skilled agents more than their higher-skilled counterparts. We study the staggered introduction of a generative AI–based conversational assistant using data from 5,172 customer-support agents. Access to AI assistance increases worker productivity, as measured by issues resolved per hour, by 15% on average, with substantial heterogeneity across workers. The effects vary significantly across different agents. Less experienced and lower-skilled workers improve both the speed and quality of their output, while the most experienced and highest-skilled workers see small gains in speed and small declines in quality. We also find evidence that AI assistance facilitates worker learning and improves English fluency, particularly among international agents...	Erik Brynjolfsson, Danielle Li, Lindsey Raymond
9	2024	Artificial intelligence and the skill premium: A numerical analysis of theoretical models ↗ This paper directly addresses the project's core theme of distributional effects by analyzing how AI impacts the skill premium and labor income inequality using a task-based theoretical framework. It specifically investigates whether AI substitutes or complements different skill levels, which is central to understanding the winners and losers in the AI-driven labor market. As a new engine in guiding China's high-quality economic development, it is important to study whether the development of artificial intelligence (AI) will increase the skill premium and affect labor income inequality. Based on Acemoglu and Restrepo's (2018a) task-based model, this study constructs a multi-sector dynamic general equilibrium (DGE) model to analyze the impact and mechanism of AI on the skill premium and performs a numerical simulation using China's industrial panel data from 2010 to 2019. The results show that AI widens the skill premium by substituting low-skilled labor with industrial robots and performing high-skilled labor tasks. The mechanism analysis reveals that AI...	Can Cheng, Jiayu Luo, Chun Zhu, et al.
9	2026	Navigating the Jagged Technological Frontier: Field Experimental Evidence of the Effects of Artificial Intelligence on Knowledge Worker Productivity and Quality ↗ This paper directly addresses the core question of whether AI tools augment or substitute for labor by providing field experimental evidence on AI's causal effects on worker productivity and task quality. It further contributes to the project by introducing the 'jagged technological frontier,' which highlights how AI's impact varies significantly across different types of knowledge-intensive tasks. We introduce and study the concept of a “jagged technology frontier” to describe the uneven impact of artificial intelligence (AI) capabilities, where AI assistance improves performance for some tasks but worsens it for others, even within the same knowledge workflow and with a seemingly similar level of difficulty. In collaboration with the global management consulting firm Boston Consulting Group, we have developed realistic management consulting tasks and examined the human performance implications of using AI to perform complex and knowledge-intensive work. The preregistered experiment involved 758 knowledge workers. After establishing a performance baseline on similar tasks, subjects...	Fabrizio Dell’Acqua, Edward McFowland, Ethan Mollick, et al.
9	2025	AI innovation and the labor share in European regions ↗ This paper directly addresses the project's core question regarding the causal effects of AI on wages and inequality, specifically examining the distributional consequences between capital and labor. It provides relevant empirical evidence on how AI adoption differentially impacts workers by skill level, aligning with the project's focus on winners and losers in the labor market. This paper examines how the development of Artificial Intelligence (AI) affects the distribution of income between capital and labor, and how these shifts contribute to regional income inequality. To investigate this issue, we analyze data from European regions dating back to 2000. We find that for every doubling of regional AI innovation, the labor share declines by 0.5% to 1.6%, potentially reducing it by 0.09 to 0.31 percentage points from an average of 52%, solely due to AI. This new technology has a particularly negative impact on high- and medium-skill workers, primarily through wage compression, while for low-skill workers, employment expansion induced by AI mildly offsets the...	Antonio Minniti, Klaus Prettner, Francesco Venturini
9	2020	Digitization-based automation and occupational dynamics ↗ This paper directly addresses the project's core theme of measuring AI exposure and its causal effects on employment and wages by examining historical automation risks and their labor market outcomes. It provides key empirical context on how automation reshapes occupational dynamics and creates heterogeneity in economic impacts across skill levels. We examine the relationship between occupational automation probabilities and employment dynamics over nearly two decades. We show that employment and wage shares of occupations with a higher automation risk have declined in Sweden over the period 1996-2013. This has occurred both at the aggregate private business sector but also within firms, where the wage share changes have been larger than the employment share changes. Combining the automation risk in workers’ occupations with individual worker characteristics, we find substantial heterogeneity. This includes that education dampens the automation risk of workers, as the average automation probability of low-skilled workers is almost...	Malin Gardberg, Fredrik Heyman, Pehr-Johan Norbäck, et al.
9	2023	The Turing Transformation: Artificial Intelligence, Intelligence Augmentation, and Skill Premiums ↗ This paper directly addresses the project's core questions by analyzing how AI automation impacts skill premiums, inequality, and labor market distributional effects through the lens of task-based augmentation. It provides a theoretical framework for understanding whether AI acts as a substitute or complement to labor, which is central to the researcher's investigation into winner/loser dynamics and aggregate wage effects. We ask whether a technical objective of using human performance of tasks as a benchmark for AI performance will result in the negative outcomes highlighted in prior work in terms of jobs and inequality. Instead, we argue that task automation, especially when driven by AI advances, can enhance job prospects and potentially widen the scope for employment of many workers. The neglected mechanism we highlight is the potential for changes in the skill premium where AI automation of tasks exogenously improves the value of the skills of many workers, expands the pool of available workers to perform other tasks, and, in the process, increases labor income and potentially reduces inequality. We...	Ajay Agrawal, Joshua S. Gans, Avi Goldfarb
9	2025	Winners and losers of generative AI: Early Evidence of Shifts in Freelancer Demand ↗ This paper directly addresses the project's core questions regarding how AI substitutes or complements labor, identifying specific winners and losers among freelancers by skill type. It provides empirical evidence on task reorganization and the impact on entry-level workers, aligning closely with the themes of AI exposure measurement and distributional effects. We examine how ChatGPT has changed the demand for freelancers in jobs where generative AI tools can act as substitutes or complements to human labor. Using BERTopic we partition job postings from a leading online freelancing platform into 116 fine-grained skill clusters and with GPT-4o we classify them as substitutable, complementary or unaffected by LLMs. Our analysis reveals that labor demand increased after the launch of ChatGPT, but only in skill clusters that were complementary to or unaffected by the AI tool. In contrast, demand for substitutable skills, such as writing and translation, decreased by 20–50% relative to the counterfactual trend, with the sharpest decline observed for...	Ole Teutloff, Johanna Einsiedler, otto kässi, et al.
9	2022	How Are Patented AI, Software and Robot Technologies Related to Wage Changes in the United States? ↗ This paper directly addresses the project's core question regarding the causal effects of AI on wages and inequality by analyzing individual-level wage changes in response to patented AI, software, and robot technologies. It provides empirical evidence on whether AI acts as a complement or substitute for labor, finding positive wage effects that align with the investigation into distributional outcomes and skill complementarity. We analyze the relationships of three different types of patented technologies, namely artificial intelligence, software and industrial robots, with individual-level wage changes in the United States from 2011 to 2021. The aim of the study is to investigate if the availability of AI technologies is associated with increases or decreases in individual workers' wages and how this association compares to previous innovations related to software and industrial robots. Our analysis is based on available indicators extracted from the text of patents to measure the exposure of occupations to these three types of technologies. We combine data on individual wages for the United States with the new...	Frank M. Fossen, Daniel Samaan, Alina Sorgner
9	2024	The impact of AI on the workforce: Tasks versus jobs? ↗ This paper directly addresses the core question of whether AI substitutes for or augments labor by providing empirical evidence on the differential impacts on tasks versus jobs. It offers crucial insights into the mechanisms of AI adoption and its immediate effects on firm-level employment, which is central to understanding the distributional and aggregate labor market effects of AI. Will the adoption of AI by businesses substitute for worker tasks or jobs? This is a core question for which relatively scarce evidence exists—especially in the wake of recent advances in generative AI. Using a new large-scale business survey by the U.S. Census Bureau, we find that AI use is having a much greater impact on worker tasks than on employment levels at the firm level. About 27% of firms using AI report replacing worker tasks, but only about 5% experience employment change due to AI use. These rates are expected to increase to nearly 35% and 12%, respectively, in the near future.	Kathryn R. Bonney, Cory Breaux, Catherine Buffington, et al.
9	2023	'Generate' the Future of Work through AI: Empirical Evidence from Online Labor Markets ↗ This paper directly addresses the project's core questions by empirically analyzing the causal effects of Generative AI on employment and wages using an online labor market context. It specifically investigates the substitution versus augmentation dynamics for workers and identifies distributional winners and losers, aligning closely with the research themes of AI exposure measurement and task-based labor market impacts. With the advent of general-purpose Generative AI, the interest in discerning its impact on the labor market escalates. In an attempt to bridge the extant empirical void, we interpret the launch of ChatGPT as an exogenous shock, and implement a Difference-in-Differences (DID) approach to quantify its influence on text-related jobs and freelancers within an online labor marketplace. Our results reveal a significant decrease in transaction volume for gigs and freelancers directly exposed to ChatGPT. Additionally, this decline is particularly marked in units of relatively higher past transaction volume or lower quality standards. Yet, the negative effect is not universally experienced among...	Jin Liu, Xingchen Xu, Yongjun Li, et al.
9	2023	The Short-Term Effects of Generative Artificial Intelligence on Employment: Evidence from an Online Labor Market ↗ This paper directly addresses the causal effects of generative AI on employment and wages by analyzing an online labor market, which is a core question of the project. It provides empirical evidence on how AI impacts different worker groups, specifically examining the substitution effects and distributional consequences for freelancers. Generative artificial intelligence (AI) holds the potential to either complement workers by enhancing their productivity or substitute them. We examine the short-term effects of the recently released generative AI models (ChatGPT, DALL-E 2, and Midjourney) on the employment outcomes of freelancers on a large online platform. We find that freelancers in highly affected occupations suffer from the introduction of generative AI, experiencing reductions in both employment and earnings. We find similar effects studying the release of other image-based generative AI models. Exploring the heterogeneity by freelancers’ employment history, we do not find evidence that high-quality service, measured...	Xiang Hui, Oren Reshef, Luofeng Zhou
9	2024	The potential impact of AI innovations on US occupations ↗ This paper directly addresses the project's core theme of measuring AI exposure by developing a precise, task-based metric using deep learning and patent data. It provides critical insights into which occupations and skills are affected, distinguishing between augmentation and substitution, which is central to the research's investigation of labor market impacts. An occupation is comprised of interconnected tasks, and it is these tasks, not occupations themselves, that are affected by Artificial Intelligence (AI). To evaluate how tasks may be impacted, previous approaches utilized manual annotations or coarse-grained matching. Leveraging recent advancements in machine learning, we replace coarse-grained matching with more precise deep learning approaches. Introducing the AI Impact measure, we employ Deep Learning Natural Language Processing to automatically identify AI patents that may impact various occupational tasks at scale. Our methodology relies on a comprehensive dataset of 17,879 task descriptions and quantifies AI's potential impact through...	Ali Akbar Septiandri, Marios Constantinides, Daniele Quercia
9	2023	AI-Driven Labor Substitution: Evidence from Google Translate and ChatGPT ↗ This paper directly addresses the project's core question of whether AI substitutes for labor by providing causal evidence from online labor markets. It specifically examines heterogeneous effects on task types, aligning with the research focus on task-based frameworks and the distinction between augmenting versus substituting AI impacts. Although artificial intelligence (AI) has the potential to significantly disrupt businesses across a range of industries, we have limited empirical evidence for its substitution effect on human labor. We use Google’s introduction of neural network-based translation (GNNT) in 2016-2017 as a natural experiment to examine the substitution of human translators by AI in the context of a large online labor market. Using a difference-indifferences design, we show that the introduction of GNNT reduced the number of (human translation) transactions at both the overall market and individual translator levels. In addition, we show that GNNT had a stronger effect on translation tasks with analytical...	Erdem Dogukan Yilmaz, Ivana Naumovska, Vikas A. Aggarwal
9	2022	Modelling artificial intelligence in economics ↗ This paper directly addresses the core question of whether AI tools augment or substitute for labor by modeling the interaction between AI abilities and human skills. It provides a theoretical framework for understanding the distributional effects of AI on labor income, which is a central theme of the research project. Abstract We provide a partial equilibrium model wherein AI provides abilities combined with human skills to provide an aggregate intermediate service good. We use the model to find that the extent of automation through AI will be greater if (a) the economy is relatively abundant in sophisticated programs and machine abilities compared to human skills; (b) the economy hosts a relatively large number of AI-providing firms and experts; and (c) the task-specific productivity of AI services is relatively high compared to the task-specific productivity of general labor and labor skills. We also illustrate that the contribution of AI to aggregate productive labor service depends not only on the...	Thomas Gries, Wim Naudé
9	2025	Complement or Substitute? How AI Increases the Demand for Human Skills ↗ This paper directly addresses the core question of whether AI augments or substitutes labor by empirically demonstrating that AI adoption increases the demand and wage premiums for complementary human skills. It provides crucial evidence on distributional effects and skill complementarity, which are central themes in the researcher's project. Artificial Intelligence (AI) is transforming the nature of work, yet there is limited empirical evidence on how it affects demand for human skills. This paper examines whether AI adoption increases the prevalence and value of human capabilities that complement technical AI skills, such as analytical thinking, resilience, or ethical judgment, within and beyond AI-intensive job roles. Using a dataset of nearly 30 million job postings from the US, the UK and Australia, between 2018 and 2024, we distinguish between internal effects (within AI roles) and external effects (in non-AI roles) across companies, industries, and regions. This paper has three main findings. First, we find that...	Elina Mäkelä, Fabian Stephany
9	2023	Task-interdependencies between Generative AI and Workers ↗ This paper directly addresses the project's core interest in how AI tools augment or substitute for labor by formalizing task-interdependencies between workers and Generative AI. It provides microeconomic foundations for understanding firm reorganization and productivity effects, which are central to the researcher's inquiry into AI's causal effects on labor markets. Our paper formalizes a production function to give microeconomic foundations for the adoption of Generative AI (GAI) within workplaces. The production function accounts for task-interdependencies, the worker-GAI interaction and indistinguishability between human-created and AI-generated outputs. We show that workers and GAI represent two distinct but interdependent sides of the production, that jointly generate a network externality in learning that drives productivity. We find that in open learning organizations favoring the worker-GAI interaction, GAI should be matched to workers based on their ability to detect errors. We analyze configurations where the worker-GAI interaction is...	Emmanuelle Walkowiak
9	2021	Threatened by AI: Analyzing Users’ Responses to the Introduction of AI in a Crowd-sourcing Platform ↗ This paper directly addresses the project's core questions regarding how AI exposure reshapes labor markets by examining worker responses in an online platform. It provides empirical evidence on task reorganization, skill complementarity versus substitution, and the distributional effects of AI on entry-level versus skilled workers. As artificial intelligence (AI) solutions are being rapidly deployed, they increasingly compete with human labor. This study examines designers’ strategies in response to the threat from the introduction of an AI system for simple logo designs in a crowdsourcing design platform. We find that, although designers with lower abilities are more likely to exit the platform, designers with higher abilities move away from the locus of threat in the lower-tier contests and switch to more-complex design contests after the introduction of the AI system. More interestingly, we find that, although unsuccessful designers respond to the threat from AI by increasing their participation across multiple...	Mikhail Lysyakov, Siva Viswanathan
9	2024	The Employment Impact of Emerging Digital Technologies ↗ This paper directly addresses the project's core questions on how digital technology exposure varies across occupations and its causal effects on employment and inequality. It provides key empirical evidence on distributional impacts by skill level and age, fitting the themes of AI exposure measurement and aggregate labor market effects. This paper estimates the exposure of US occupations and industries to emerging digital technologies and their impact on US commuting zone (CZ) employment. Building upon the natural language processing approach introduced by Prytkova et al. (2024), we estimate the exposure of O ⋆ NET-SOC occupations and NAICS industries, thereby extending the open–access ‘TechXposure’ database to the US context. Using this new data source, we apply a shift-share design to instrument the CZ exposure to emerging digital technologies and estimate their employment impact across CZs between 2012 and 2019. We find that digital technologies have an overall positive net impact on US employment. However, the impact...	Ekaterina Prytkova, Fabien Petit, Deyu Li, et al.
9	2020	ARTIFICIAL INTELLIGENCE AND HUMAN JOBS ↗ This paper directly addresses the project's core question regarding the causal effects of AI on employment and wages, specifically examining the distributional impacts between skilled and unskilled workers. It contributes to the theme of whether AI acts as a substitute or complement to labor by modeling how AI adoption influences overall unemployment and labor market outcomes across different skill levels. The development of artificial intelligence (AI) does influence human jobs but not necessarily in a negative way. Although labor force participation rates and firms’ job vacancies for human labor decline, the unemployment rate may be lower than that in an economy without AI. In an economy with heterogeneously skilled workers, the invention of AI usually has a negative effect on the skilled labor market but a positive effect on the unskilled labor market. The overall unemployment rate may decline as AI develops.	Chia‐Hui Lu
9	2025	Artificial intelligence and the skill premium ↗ This paper directly addresses the project's core question regarding whether AI acts as a complement or substitute for labor and its distributional effects on skill premiums. It explicitly models the differential impact of AI on high-skill versus low-skill workers, providing theoretical insight into the mechanisms of AI-driven inequality. How will ChatGPT and other forms of artificial intelligence (AI) affect the skill premium? To address this question, we propose a nested constant elasticity of substitution production function that distinguishes among three types of capital: traditional physical capital (machines, assembly lines), industrial robots, and AI. Following the literature, we assume that industrial robots predominantly substitute for low-skill workers, whereas AI mainly helps to perform the tasks of high-skill workers. We show that AI reduces the skill premium as long as it is more substitutable for high-skill workers than low-skill workers are for high-skill workers.	David E. Bloom, Klaus Prettner, Jamel Saadaoui, et al.
9	2023	Measuring the Impact of AI on Information Worker Productivity ↗ [Title only] The title directly addresses the project's core theme of generative AI productivity experiments and measures the causal effect of AI on worker productivity. It aligns perfectly with the investigation into whether AI tools augment or substitute for labor in information-based occupations. No abstract available.	Benjamin Edelman, Donald Ngwe, Sida Peng
9	2024	Artificial Intelligence, Scientific Discovery, and Product Innovation ↗ This paper directly addresses core project themes by providing empirical evidence on how AI augments labor productivity and the complementary role of human expertise in innovative tasks. It further contributes by analyzing distributional effects across skill levels and the reorganization of work, specifically noting the trade-offs between efficiency gains and worker satisfaction. This paper studies the impact of artificial intelligence on innovation, exploiting the randomized introduction of a new materials discovery technology to 1,018 scientists in the R&D lab of a large U.S. firm. AI-assisted researchers discover 44% more materials, resulting in a 39% increase in patent filings and a 17% rise in downstream product innovation. These compounds possess more novel chemical structures and lead to more radical inventions. However, the technology has strikingly disparate effects across the productivity distribution: while the bottom third of scientists see little benefit, the output of top researchers nearly doubles. Investigating the mechanisms behind these results, I...	Aidan Toner-Rodgers
9	2023	Artificial Intelligence and Jobs: Evidence from US Commuting Zones ↗ This paper directly addresses the project's core question regarding the causal effects of AI on employment and inequality by providing empirical evidence from US commuting zones. It specifically examines the distributional winners and losers by skill level and occupation, aligning perfectly with the themes of AI exposure measurement and aggregate labor market effects. We study the effect of Artificial Intelligence (AI) on employment across US commuting zones over the period 2000-2020. A simple model shows that AI can automate jobs or complement workers, and illustrates how to estimate its effect by exploiting variation in a novel measure of local exposure to AI: job growth in AI-related professions built from detailed occupational data. Using a shift-share instrument that combines industry-level AI adoption with local industry employment, we estimate robust negative effects of AI exposure on employment across commuting zones and time. We find that AI’s impact is different from other capital and technologies, and that it works through services more than...	Alessandra Bonfiglioli, Gino Gancia, Ioannis Papadakis, et al.
9	2025	Roles of Artificial Intelligence in Collaboration with Humans: Automation, Augmentation, and the Future of Work ↗ This paper directly addresses the core project theme of whether AI tools augment or substitute for labor by distinguishing between automation and augmentation roles. It provides a theoretical and empirical framework for task reorganization and human-AI collaboration, which is central to understanding how AI reshapes labor markets and worker productivity. Humans will see significant changes in the future of work as collaboration with artificial intelligence (AI) will become commonplace. This work explores the benefits of AI in the setting of judgment tasks when it replaces humans (automation) and when it works with humans (augmentation). Through an analytical modeling framework, we show that the optimal use of AI for automation or augmentation depends on different types of human-AI complementarity. Our analysis demonstrates that the use of automation increases with higher levels of between-task complementarity. In contrast, the use of augmentation increases with higher levels of within-task complementarity. We integrate both automation and...	Andreas Fügener, Dominik D. Walzner, Alok Gupta
9	2024	Is this time different? Impact of AI in output, employment and inequality across low, middle and high-income countries ↗ This paper directly addresses the project's core questions regarding economywide employment effects and distributional impacts of AI across different income groups using a macroeconomic simulation framework. It specifically investigates the tension between labor substitution and innovation-driven job creation, providing crucial context on aggregate labor market outcomes and inequality. Recent developments in artificial intelligence (AI) have revived concerns about technological unemployment and the increase in inequality due to technological change. Many studies of the impact of automation have considered a static picture of economies that put less emphasis on their potential for job creation through product innovation. They also tend to focus on developed countries and do not consider the potential impact on developing countries through trade and changes in specialisation patterns. This paper studies the impact of AI on the GDP and employment in low, middle and high-income countries, as well as on the income inequality across countries, based on computer simulations of a...	Clovis Freire
9	2024	The Productivity Effects of Generative AI: Evidence from a Field Experiment with GitHub Copilot ↗ This paper directly addresses the project's core theme of generative AI productivity experiments by providing field evidence from software developers. It offers crucial empirical data on how AI tools like GitHub Copilot affect worker output, which is central to understanding the causal effects of AI on productivity and task augmentation. We are providing a preview of a project that analyzes two field experiments with 1,974 software developers at Microsoft and Accenture to evaluate the productivity impact of Generative AI. As part of our study, a random subset of developers was given access to GitHub Copilot, an AI-based coding assistant that intelligently suggests âcompletionsâ for code. Our preliminary results provide suggestive evidence that these developers became more productive, completing 12.92% to 21.83% more pull requests per week at Microsoft and 7.51% to 8.69% at Accenture (depending on specification). Due to low compliance in the Microsoft experiment and internal organizational changes at Accenture, our...	Kevin Zheyuan Cui, Mert Demirer, Sonia Jaffe, et al.
9	2023	Generative AI, generating precariousness for workers? ↗ [Title only] This title directly addresses the distributional effects of generative AI, specifically focusing on labor market instability and inequality, which are core themes of the project. The use of 'precariousness' suggests an analysis of who the losers are, aligning with the investigation into winners, losers, and the reshaping of work conditions by AI. No abstract available.	Aida Ponce del Castillo
9	2023	Skills or Degree? The Rise of Skill-Based Hiring for AI and Green Jobs ↗ This paper directly addresses the project's themes of AI exposure measurement, skill complementarity vs. substitution, and distributional effects by analyzing how skill-based hiring practices reshape recruitment and wage premiums in AI occupations. It provides empirical evidence on how firms reorganize labor demand by prioritizing specific technical skills over formal degrees, highlighting winners and losers in the emerging AI labor market. Emerging professions in fields like Artificial Intelligence (AI) and sustainability (green jobs) are experiencing labour shortages as industry demand outpaces labour supply. In this context, our study aims to understand whether employers have begun focusing more on individual skills rather than formal qualifications in their recruitment processes. We analysed a large time-series dataset of approximately eleven million online job vacancies in the UK from 2018 to mid-2024, drawing on diverse literature on technological change and labour market signalling. Our findings provide evidence that employers have initiated"skill-based hiring"for AI roles, adopting more flexible hiring practices to...	Eugenia González Ehlinger, Fabian Stephany
9	2024	The Effects of Generative AI on High Skilled Work: Evidence from Three Field Experimentswith Software Developers ↗ [Title only] This paper directly addresses the core themes of generative AI productivity experiments and their impact on high-skilled workers, which is central to understanding augmentation vs. substitution. It provides causal evidence on how AI affects a specific occupation, helping to identify winners and losers within the labor market. No abstract available.	sida peng
9	2024	Automated machines and the labor wage gap ↗ This paper directly addresses the project's core themes by employing a task-based framework to analyze how automation alters labor wage gaps and capital-labor complementarity. It provides empirical evidence on the distributional effects of technological change, specifically focusing on how different types of capital investment impact wages for complementary labor, which aligns with the investigation into AI's substitution versus augmentation effects. This paper analyzes how the increase in the proportion of automated machines to general capital inputs affects the wage gap of the labor force complementary to the two types of capital, the general equilibrium model and constant elasticity of substitution production function are introduced and the impact mechanism is explored using a task-based model. Data for listed Chinese manufacturing companies from 2012 to 2020 is used and a double fixed effects model is invoked for empirical testing. The findings show that (1) a rise in the proportion of automated machines to general capital inputs increases the wage gap of labor complementary to automated machine capital and general capital, and a...	Hong Jiang, Xue Wang, Chongguang Liu
9	2025	How Different Uses of AI Shape Labor Demand: Evidence from France ↗ This paper directly addresses the project's core questions regarding how AI reshapes labor markets by providing causal evidence on firm-level employment and productivity effects. It specifically investigates whether AI acts as a complement or substitute across different tasks and uses, identifying distinct winners and losers based on occupational classification. Using French firm-level data on AI adoption from 2017-2020, we find that, first, firms adopting AI are larger and more productive and skill intensive. Second, difference-in-difference estimates reveal an increase in firm-level employment and sales after AI adoption, suggesting that the induced productivity gains allow firms to grow and outweigh potential displacement effects. Third, occupations classified in recent work as substitutable with AI expand. Fourth, AI usage is a relevant dimension of heterogeneity in the labor demand response: We find positive employment growth for certain uses (e.g., information and communications technology security) and negative for others (e.g...	Philippe Aghion, Simon Bunel, Xavier Jaravel, et al.
9	2023	AI and Jobs: Has the Inflection Point Arrived? Evidence from an Online Labor Platform ↗ This paper directly addresses the project's core questions by analyzing how generative AI augments or substitutes for labor across different occupations using online labor market data. It provides empirical evidence on heterogeneous effects by skill and region, while introducing a theoretical framework to identify inflection points where AI shifts from enhancement to displacement. This study investigates how artificial intelligence (AI) influences various online labor markets (OLMs) over time. Employing the Difference-in-Differences method, we discovered two distinct scenarios following ChatGPT's launch: displacement effects featuring reduced work volume and earnings, exemplified by translation&localization OLM; productivity effects featuring increased work volume and earnings, exemplified by web development OLM. To understand these opposite effects in a unified framework, we developed a Cournot competition model to identify an inflection point for each market. Before this point, human workers benefit from AI enhancements; beyond this point, human workers would be...	Dandan Qiao, Huaxia Rui, Xiong Qian
9	2023	Automation: Theory, Evidence, and Outlook ↗ This paper directly aligns with the project's core task-based framework by reviewing the theoretical and empirical evidence on how automation substitutes for labor. It addresses key questions regarding the causal effects of technology on employment, wages, and productivity, serving as a foundational overview for understanding the mechanisms behind AI's impact on labor markets. This article reviews the literature on automation and its impact on labor markets, wages, factor shares, and productivity. I first introduce the task model and explain why this framework offers a compelling way to think about recent labor market trends and the effects of automation technologies. The task model clarifies that automation technologies operate by substituting capital for labor in a widening range of tasks. This substitution reduces costs, creating a positive productivity effect, but also reduces employment opportunities for workers displaced from automated tasks, creating a negative displacement effect. I survey the empirical literature and conclude that there is wide...	Pascual Restrepo
9	2024	Artificial Intelligence and the Skill Premium ↗ This paper directly addresses the project's core question regarding the distributional effects of AI on wages and inequality, specifically focusing on the skill premium. It explicitly investigates whether AI acts as a complement or substitute for different skill levels, which is central to understanding the winners and losers of AI adoption. How will the emergence of ChatGPT and other forms of artificial intelligence (AI) affect the skill premium?To address this question, we propose a nested constant elasticity of substitution production function that distinguishes among three types of capital: traditional physical capital (machines, assembly lines), industrial robots, and AI.Following the literature, we assume that industrial robots predominantly substitute for low-skill workers, whereas AI mainly helps to perform the tasks of high-skill workers.We show that AI reduces the skill premium as long as it is more substitutable for high-skill workers than low-skill workers are for high-skill workers.	David E. Bloom, Klaus Prettner, Jamel Saadaoui, et al.
9	2023	Technology and Labor Displacement: Evidence from Linking Patents with Worker-Level Data ↗ This paper directly addresses the project's core themes by measuring technology exposure at the task level and analyzing its causal effects on worker earnings and employment. It specifically contributes to understanding the distributional impacts of technology on workers by skill and age, and explicitly extends its findings to predict the effects of AI. We develop measures of labor-saving and labor-augmenting technology exposure using textual analysis of patents and job tasks.Using US administrative data, we show that both measures negatively predict earnings growth of individual incumbent workers.While labor-saving technologies predict earnings declines and higher likelihood of job loss for all workers, laboraugmenting technologies primarily predict losses for older or highly-paid workers.However, we find positive effects of labor-augmenting technologies on occupation-level employment and wage bills.A model featuring labor-saving and labor-augmenting technologies with vintage-specific human capital quantitatively matches these patterns.We...	Leonid Kogan, Dimitris Papanikolaou, Lawrence Schmidt, et al.
9	2025	Human + AI in Accounting: Early Evidence from the Field ↗ This paper directly addresses the project's core themes by providing field evidence on how generative AI augments rather than substitutes labor in accounting tasks. It offers specific insights into task reorganization, productivity gains, and the complementarity between professional expertise and AI tools. ABSTRACT This paper provides early evidence on the integration and impact of generative artificial intelligence (GenAI) in accounting at the accountant and task levels. Using survey data from 277 professional accountants, we document substantial heterogeneity in adoption patterns, perceived benefits, and concerns about GenAI. Using proprietary field data from an AI‐enabled accounting platform serving 79 small‐ and medium‐sized enterprises, we analyze over 200,000 transaction‐level records. We document that GenAI adoption is associated with significant productivity gains and systematic reallocation of effort away from routine data entry toward business communication and quality assurance...	Jung Ho Choi, Chloe Xie
9	2024	Employer and Employee Responses to Generative AI: Early Evidence ↗ [Title only] This title directly addresses the core themes of AI exposure, firm adoption, and distributional effects by focusing on the early responses from both employers and employees. It likely provides crucial empirical evidence on how generative AI is reshaping labor markets and worker-firm dynamics, which aligns perfectly with the project's investigation into causal effects and task reorganization. No abstract available.	Philip G. Berger, Wei Cai, Lin Qiu, et al.
9	2024	How effective is AI augmentation in human–AI collaboration? Evidence from a field experiment ↗ This paper directly addresses the project's core question of whether AI tools augment or substitute for labor by providing causal evidence from a field experiment showing AI augmentation increases productivity. It also contributes to understanding distributional effects and the fate of entry-level workers by highlighting that inexperienced agents benefit significantly more than experienced ones. Purpose Companies increasingly leverage artificial intelligence (AI) to enhance human performance, particularly in e-commerce. However, the effectiveness of AI augmentation remains controversial. This study investigates whether, how and why AI enhances human agents’ sales through a randomized field experiment. Design/methodology/approach This study conducts a two-by-two factorial randomized field experiment ( N = 1,090) to investigate the effects of AI augmentation on sales. The experiment compares sales outcomes handled solely by human agents with those augmented by AI, while also examining the moderating effect of agents’ experience levels and the underlying mechanisms behind agents’...	Chengcheng Liao, Xin Wen, Shan Li, et al.
9	2024	Complement or substitute? How AI increases the demand for human skills ↗ This paper directly addresses the core project theme of whether AI augments or substitutes for labor by providing empirical evidence on the complementarity between AI and specific human skills. It offers crucial insights into distributional effects and skill demand shifts, which are central to understanding the causal impacts of AI on wages and employment structures. Artificial Intelligence (AI) is transforming the nature of work, yet there is limited empirical evidence on how it affects demand for human skills. This paper examines whether AI adoption increases the prevalence and value of human capabilities that complement technical AI skills, such as analytical thinking, resilience, or ethical judgment, within and beyond AI-intensive job roles. Using a dataset of nearly 30 million job postings from the US, the UK and Australia, between 2018 and 2024, we distinguish between internal effects (within AI roles) and external effects (in non-AI roles) across companies, industries, and regions. This paper has three main findings. First, we find that...	Elina Mäkelä, Fabian Stephany, Sehrer, Mareike, et al.
9	2024	The EPOCH of AI: Human-Machine Complementarities at Work ↗ This paper directly addresses the project's core questions regarding AI exposure measurement and the distinction between labor augmentation and substitution through a novel theoretical framework. It provides empirical evidence on how these distinct AI roles impact employment trends and skill demands across occupations, aligning closely with the research's focus on distributional effects and task-based analysis. We introduce the EPOCH framework (Empathy, Presence, Opinion, Creativity, and Hope) to capture human capabilities that complement, rather than substitute, artificial intelligence. Using network-based methods that map task interdependencies across all U.S. occupations, we develop three metrics: (i) an EPOCH score measuring human-intensive skills, (ii) a potential-for-augmentation score, and (iii) a risk-of-substitution score. This framework explicitly distinguishes AI’s roles in augmenting versus automating work, addressing a key gap in the literature. Our results show a clear shift toward more human-intensive work. New tasks emerging in 2024 carry significantly higher EPOCH scores than...	Isabella Loaiza, Roberto Rigobón
9	2025	Artificial intelligence, tasks, skills, and wages: Worker-level evidence from Germany ↗ This paper directly addresses core themes by measuring AI exposure at the worker and occupational levels within a task-based framework, distinguishing AI from traditional automation technologies. It provides empirical evidence on the causal links between AI exposure, task reorganization, skill complementarity, and wage outcomes, which are central to the project's inquiry into distributional effects and labor market reshaping. This paper examines how new technologies are linked to changes in the content of work and individual wages. As a first step, it documents novel facts on task and skill changes within occupations over the past two decades in Germany. We furthermore reveal a distinct relationship between ex-ante occupational work content and ex-post exposure to artificial intelligence (AI) and automation (robots). Workers in occupations with high AI exposure perform different activities and face different skill requirements compared to workers in occupations exposed to robots, suggesting that robots and AI are substitutes for different activities and skills. We also document that changes in the task and skill...	Erik Engberg, Michael Koch, Magnus Lodefalk, et al.
9	2025	Impact of robots and artificial intelligence on labor and skill demand: evidence from the UK ↗ This paper directly addresses the core question of how AI exposure varies by skill level and its causal effects on labor demand and wages. It provides critical empirical evidence on whether AI substitutes for high-skilled labor, a key theme regarding distributional effects and skill complementarity versus substitution. Abstract Over the past four decades, automation technologies have replaced routine tasks performed by medium-skilled workers, and contributed to increased labor market polarization. With the advent of artificial intelligence, this dynamic may have shifted, extending task substitution to non-routine tasks performed by high-skilled workers. Using textual analysis and descriptions of technology found in patent texts, we construct novel occupational exposures to robot and artificial intelligence technologies. These occupational exposures are then used to analyze changes in labor and skill demand over the last decade in the United Kingdom. We find that the middle part of the income distribution...	Martin Lábaj, Tomáš Oleš, Gabriel Procházka
9	2024	A task-based approach to inequality ↗ This paper directly addresses the project's core themes by employing a task-based framework to analyze how automation and technology adoption reshape labor markets and inequality. It provides essential theoretical context for understanding whether AI augments or substitutes for labor and how this affects the distribution of wages and employment across different skill levels. Abstract This article reviews recent work on how automation and task displacement have contributed to labour share declines and inequality in the US labour market. We summarize the basic building blocks of a task-based framework in which a set of tasks is allocated between capital, skilled labour and unskilled labour. Automation, which corresponds to the use of new technologies expanding the set of tasks that can be performed by capital, always reduces the labour share in value added and may depress overall wages and employment. The negative effects of automation on labour share and its potentially adverse consequences for labour demand can be counterbalanced by the creation of new...	Daron Acemoğlu, Pascual Restrepo
9	2025	This time it’s different – Generative artificial intelligence and occupational choice ↗ This paper directly addresses the project's core questions regarding AI's impact on labor markets by providing causal evidence on how generative AI reshapes occupational choice and entry-level worker supply. It highlights distributional effects by showing that high-language tasks are significantly affected and examines how worker search behavior and applicant quality respond to AI exposure. • ChatGPT reduced search intensity for apprenticeship vacancies by 8% on average • Decline was strongest in jobs involving cognitive tasks and high language skills • Occupations previously considered “safe” from automation were also highly affected • Changes in search behaviour align with the most recent assessments of AI exposure • Applicant quality declined, particularly for commercial apprenticeships We show the causal influence of the launch of generative artificial intelligence (AI) in the form of ChatGPT on the search behavior of young people for apprenticeship vacancies. To estimate the short- and medium-term effects, we use a variety of methods, including a...	Daniel Göller, Christian Gschwendt, Stefan C. Wolter
9	2026	The Rapid Adoption of Generative AI ↗ This paper directly addresses the project's focus on AI exposure measurement by providing granular, nationally representative data on the speed, intensity, and distribution of generative AI adoption across the U.S. workforce. It also offers critical evidence on early productivity effects and task reorganization, which are central to understanding the causal impacts of AI on labor markets. Generative artificial intelligence (genAI) is a potentially important new technology, but its impact on the economy depends on the speed and intensity of adoption. This paper reports results from a series of nationally representative U.S. surveys of genAI use at work and at home. As of late 2024, 45% of the U.S. population age 18–64 uses genAI. Among employed respondents, 27% used genAI for work at least once in the previous week: 10% used it every workday and 17% on some but not all workdays. Relative to each technology’s first mass-market product launch, work adoption of genAI has been faster than the personal computer (PC), and overall adoption has outpaced both PCs and the internet by...	Alexander Bick, Adam Blandin, David J. Deming
9	2026	The Effects of Generative AI on High-Skilled Work: Evidence from Three Field Experiments with Software Developers ↗ This paper directly addresses the project's core interest in generative AI productivity experiments and task-based framework by providing rigorous field experiment evidence on software developers. It specifically contributes to the distributional effects theme by highlighting how productivity gains vary by skill level, showing higher benefits for less experienced developers. This study evaluates the effect of generative artificial intelligence (AI) on software developer productivity via randomized controlled trials at Microsoft, Accenture, and an anonymous Fortune 100 company. These field experiments, run by the companies as part of their ordinary course of business, provided a random subset of developers with access to an AI-based coding assistant suggesting intelligent code completions. Although each experiment is noisy and results vary across experiments, when data are combined across three experiments and 4,867 developers, our analysis reveals a 26.08% increase (standard error: 10.3%) in completed tasks among developers using the AI tool. Notably, less...	Kevin Zheyuan Cui, Mert Demirer, Sonia Jaffe, et al.
9	2024	Exposure to Artificial Intelligence and Occupational Mobility: A Cross-Country Analysis ↗ This paper directly addresses the project's core question regarding who are the winners and losers from AI by analyzing occupational mobility and wage effects for different skill levels. It provides critical empirical evidence on how AI reshapes labor markets through task-based exposure and complementarity, specifically highlighting the disruption to entry-level workers and the divergent outcomes for college versus non-college educated individuals. We document historical patterns of workers' transitions across occupations and over the life-cycle for different levels of exposure and complementarity to Artificial Intelligence (AI) in Brazil and the UK. In both countries, college-educated workers frequently move from high-exposure, low-complementarity occupations (those more likely to be negatively affected by AI) to high-exposure, high-complementarity ones (those more likely to be positively affected by AI). This transition is especially common for young college-educated workers and is associated with an increase in average salaries. Young highly educated workers thus represent the demographic group for which AI-driven structural change...	Mauro Cazzaniga
9	2019	The Hardware–Software Model: A New Conceptual Framework of Production, R&D, and Growth with AI This paper directly addresses the project's core theme of AI skill complementarity versus substitution by introducing a novel hardware-software framework that models AI as cognitive work complementary to hardware. It provides a theoretical foundation for understanding how AI reshapes labor markets, factor shares, and long-run growth, aligning closely with the project's macroeconomic and task-based inquiry. The article proposes a new conceptual framework for capturing production, R&D, and economic growth in aggregative models which extend their horizon into the digital era. Two key factors of production are considered: hardware, including physical labor, traditional physical capital and programmable hardware, and software, encompassing human cognitive work, pre-programmed software, and artificial intelligence (AI). Hardware and software are complementary in production whereas their constituent components are mutually substitutable. The framework generalizes, among others, the standard model of production with capital and labor, models with capitalâ€“skill complementarity and skill-biased...	Jakub Growiec
9	2024	Artificial intelligence and wage inequality ↗ This paper directly addresses the core question of how AI affects wage inequality and distributional outcomes across workers. It utilizes an occupation-level AI exposure measure and provides empirical evidence on whether AI acts as a substitute or complement to labor, specifically examining its impact on within-occupation wage differentials. This paper looks at the links between AI and wage inequality across 19 OECD countries. It uses a measure of occupational exposure to AI derived from that developed by Felten, Raj and Seamans (2019) – a measure of the degree to which occupations rely on abilities in which AI has made the most progress. The results provide no indication that AI has affected wage inequality between occupations so far (over the period 2014-2018). At the same time, there is some evidence that AI may be associated with lower wage inequality within occupations – consistent with emerging findings from the literature that AI reduces productivity differentials between workers. Further research is needed to identify...	Alexandre Georgieff
9	2025	Collaborating with AI Agents: Field Experiments on Teamwork, Productivity, and Performance ↗ [Title only] The title explicitly mentions field experiments on productivity and performance in the context of AI collaboration, directly addressing the core theme of generative AI productivity experiments and causal effects on worker output. It likely provides empirical evidence on whether AI tools augment or substitute for labor within team settings, a key question regarding task reorganization and firm-level AI adoption. No abstract available.	Harang Ju, Sinan Aral
9	2024	The Heterogeneous Productivity Effects of Generative AI ↗ This paper directly addresses the core question of how generative AI affects worker productivity by leveraging a natural experiment to measure heterogeneous impacts across skill levels. It provides empirical evidence on task reorganization and skill complementarity versus substitution, which are central themes of the researcher's project. We analyse the individual productivity effects of Italy's ban on ChatGPT, a generative pretrained transformer chatbot. We compile data on the daily coding output quantity and quality of over 36,000 GitHub users in Italy and other European countries and combine these data with the sudden announcement of the ban in a difference-in-differences framework. Among the affected users in Italy, we find a short-term increase in output quantity and quality for less experienced users and a decrease in productivity on more routine tasks for experienced users.	David Kreitmeir, Paul A. Raschky
9	2024	Tasks at Work: Comparative Advantage, Technology and Labor Demand ↗ This paper provides the foundational task-based framework essential for analyzing how AI reshapes labor markets, directly addressing the project's core theme of measuring AI exposure and its impact on tasks. It explicitly outlines mechanisms such as automation, augmentation, and the creation of new tasks, which are central to understanding the causal effects of AI on worker productivity and employment. This chapter reviews recent advances in the task model and shows how this framework can be put to work to understand trends in the labor market in recent decades. Production in each industry requires the completion of various tasks that can be assigned to workers with different skills or to capital. Factors of production have well-defined comparative advantage across tasks, which governs substitution patterns. Technological change can: (1) augment a specific labor type—e.g., increase the productivity of labor in tasks it is already performing; (2) augment capital; (3) automate work by enabling capital to perform tasks previously allocated to labor; (4) create new tasks. The task model...	Daron Acemoğlu, Fredric Kong, Pascual Restrepo
9	2021	The Race of Man and Machine: Implications of Technology When Abilities and Demand Constraints Matter ↗ This paper directly addresses the project's core questions on how AI reshapes labor markets by proposing an endogenous growth model that analyzes the impact of AI on jobs, inequality, and wages. It contributes to the project's interest in task-based frameworks and distributional effects by modeling AI as providing 'abilities' and incorporating demand constraints to explain income inequality. IZA DP No. 14341 APRIL 2021 The Race of Man and Machine: Implications of Technology When Abilities and Demand Constraints Matter In “The Race between Man and Machine: Implications of Technology for Growth, Factor Shares, and Employment,” Acemoglu and Restrepo (2018b) combine the task-based model of the labor market with an endogenous growth model to model the economic consequences of artificial intelligence (AI). This paper provides an alternative endogenous growth model that addresses two shortcomings of their model. First, we replace the assumption of a representative household with the premise of two groups of households with different preferences. This allows our model to be demand...	Thomas Gries, Wim Naudé
9	2025	Artificial intelligence adoption and workplace training ↗ This paper directly addresses the project's core question of how firms reorganize work and adjust skill demands in response to AI adoption. It provides empirical evidence on the distributional effects of AI by showing a shift in training resources toward high-skilled workers, thereby informing the debate on whether AI acts as a substitute or complement for labor. As artificial intelligence (AI) reshapes business processes, firms must adapt their training strategies to cultivate a skilled workforce. Using German establishment-level panel data from 2019 to 2023, this study analyzes how firms adjust their training strategies following AI adoption. Staggered difference-in-differences analysis shows that sustained AI adoption is associated with a 14% increase in new apprenticeships among training firms (intensive margin), but is not linked to the training decision (extensive margin). AI adoption is also associated with a modest increase in continuing training, with resources shifting toward high-skilled employees. The results align with AI as an...	Samuel Muehlemann
9	2024	The role of firm AI capabilities in generative AI-pair coding ↗ This paper directly addresses the project's core themes of AI productivity effects, task-based reorganization, and skill complementarity within a specific high-tech occupation. It provides empirical evidence on how firm-level capabilities and organizational changes mediate the impact of generative AI on worker output, aligning closely with the research questions on augmentation versus substitution and firm adaptation. Generative Artificial Intelligence (genAI) is the latest evidence of the transformative value of AI in organizations. One promising avenue lies in software engineering, where genAI can contribute to coding by pairing with developers. Based on a sample of global firms, two main insights emerge on analyzing the productivity implications of genAI-pair coding. Coding quality is negatively correlated with productivity throughput gains, while quality-adjusted productivity gains depend on the extent to which organizations have deployed AI capabilities in the form of data, skills upgrade, and AI governance. As observed with other digital technologies, the success of using genAI is closely tied to...	Jacques Bughin
9	2020	Artificial Intelligence, Income Distribution and Economic Growth ↗ [Title only] This title directly addresses the researcher's core questions regarding the distributional effects of AI on income and its aggregate impact on economic growth. It aligns perfectly with the themes of inequality, macroeconomics of AI, and economywide effects. No abstract available.	Thomas Gries, Wim Naudé
9	2025	Star Advantage: Employee Value Creation and Capture in the Age of Artificial Intelligence ↗ The paper directly addresses the project's core questions regarding distributional effects, worker inequality, and the mechanism of AI complementarity versus substitution. It specifically analyzes how AI adoption widens performance gaps and reshapes value capture by skill level, which is central to understanding the winners and losers in the AI labor market. ABSTRACT The integration of generative artificial intelligence (AI) into knowledge work is fundamentally reshaping employee performance and value creation in ways that challenge conventional wisdom. Rather than performance disparities being reduced through AI adoption, we argue that they may increase as star employees leverage superior domain expertise and strategic AI deployment to widen performance gaps—a phenomenon we term the “AI‐specific Matthew Effect.” These performance transformations coincide with dramatic shifts in value appropriation dynamics: Personal AI tools will enhance employee bargaining power by enabling portable, high‐value outputs independent of organizational resources...	Matthew L. Call, Kaifeng Jiang, Connor Idso
9	2025	Early Impacts of M365 Copilot ↗ This paper directly addresses the project's core theme of generative AI productivity experiments by providing rigorous causal evidence from a large-scale randomized trial. It specifically examines how AI tools augment knowledge worker productivity, a key mechanism for understanding labor market impacts and task reorganization. Advances in generative AI have rapidly expanded the potential of computers to perform or assist in a wide array of tasks traditionally performed by humans. We analyze a large, real-world randomized experiment of over 6,000 workers at 56 firms to present some of the earliest evidence on how these technologies are changing the way knowledge workers do their jobs. We find substantial time savings on common core tasks across a wide range of industries and occupations: workers who make use of this technology spent half an hour less reading email each week and completed documents 12% faster. Despite the newness of the technology, nearly 40% of workers who were given access to the tool used it...	Eleanor Wiske Dillon, Sonia Jaffe, Sida Peng, et al.
9	2025	A Gender Lens on Labor Market Exposure to AI ↗ This paper directly addresses the project's core theme of distributional effects by analyzing how AI exposure and substitution risks vary by gender and wage level. It provides critical insights into the 'who are the winners and losers' question, particularly regarding vulnerability among low-earning female workers. The rise of AI may profoundly impact labor markets, as AI tools could perform numerous cognitive tasks traditionally in the human domain. This paper examines the gendered effects of AI adoption across six economies of varying income levels. In most countries, a higher share of women than men work in AI-exposed occupations, making them more likely to benefit from the technology but also more vulnerable to disruptions. Within jobs facing higher risk of substitution, women at the bottom of the wage distribution make more frequent transitions to inactivity, suggesting that AI could pose a particular threat to low-earning females.	Mauro Cazzaniga, Augustus Panton, Longji Li, et al.
9	2023	Task-Interdependencies between Generative AI and Workers ↗ [Title only] This title directly addresses the core themes of task-based frameworks and whether AI tools augment or substitute for labor by focusing on the interdependencies between generative AI and workers. It is highly relevant to understanding firm reorganization and the specific mechanisms through which AI reshapes labor markets. No abstract available.	Emmanuelle Walkowiak
9	2025	Robots & AI exposure and wage inequality: a within occupation approach ↗ This paper directly addresses the core theme of how AI exposure affects wage inequality and distributional outcomes by distinguishing between the effects of industrial robots and artificial intelligence. It provides crucial empirical evidence on whether AI acts as a complement to high-skilled labor, thereby increasing wage dispersion, which is central to understanding the labor market impacts of AI. This paper examines the linkages between occupational exposure to recent automation technologies and inequality across 19 European countries. Using data from the European Union Structure of Earnings Survey (EU-SES), a fixed-effects model is employed to assess the association between occupational exposure to artificial intelligence (AI) and to industrial robots–two distinct forms of automation–and within-occupation wage inequality. The analysis reveals that occupations with higher exposure to robots tend to have lower wage inequality, particularly among workers in the lower half of the wage distribution. In contrast, occupations more exposed to AI exhibit greater wage dispersion, especially...	Florencia Jaccoud
9	2025	AI and the labour market: opening the black box ↗ This paper directly addresses the project's core themes by providing a comprehensive review of AI exposure measurement and analyzing its heterogeneous impacts on occupations, skills, and organizational structures. It explicitly covers key questions regarding who the winners and losers are across different demographic and sectoral groups, aligning perfectly with the research focus on distributional effects and task-based frameworks. This work aims at discussing some of the main (open) questions about the labour impact of AI technologies. First, we provide an in-depth literature review focusing on concepts and measurement approaches and distinguishing between up (invention and knowledge creation), mid (technological innovation and development) and downstream (adoption and diffusion) components of the AI value chain. Second, we summarise the six articles included in the Special Issue ‘AI and labor markets: opening the black box’, distinguishing between contributions focusing on AI exposure, occupations and skill demand; the relationship between AI and automation technologies and their impact on income distribution; and...	Nathalie Greenan, Dario Guarascio, Jelena Reljic
9	2024	New Technologies: End of Work or Structural Change? ↗ This paper directly addresses the project's core questions by analyzing how AI and automation reshape labor markets through structural change, workforce composition shifts, and inequality. It provides crucial insights into the distributional effects, skill complementarity, and policy implications relevant to the research themes. Abstract This paper examines the impact of new technologies, particularly automation and artificial intelligence (AI), on labor markets. The existing literature documents ambiguous and only limited overall employment effects, while new technologies induce significant shifts in workforce composition. The implied firm-level productivity gains primarily benefit larger, skilled-labor-intensive firms. AI adoption remains limited but continues to reshape skill demands. The implied worker reallocation is costly, exacerbating inequality. This calls for policies such as targeted support for displaced workers, investment in education and skill development, promoting technology diffusion, and...	Emilie Rademakers, Ulrich Zierahn
9	2025	AI, Task Changes in Jobs, and Worker Reallocation ↗ This paper directly addresses the project's core themes by examining how AI reshapes task content within occupations and the resulting worker reallocation patterns. It provides empirical evidence on the causal effects of AI on wages and employment, specifically highlighting the distributional impacts between low-skilled and high-skilled workers. How does Artificial Intelligence (AI) affect the task content of work, and how do workers adjust to the diffusion of AI in the economy? To answer these important questions, we combine novel patent-based measures of AI and robot exposure with individual survey data on tasks performed on the job and administrative data on worker careers. Like prior studies, we find that robots have reduced routine tasks. In sharp contrast, AI has reduced non-routine abstract tasks like information gathering and increased the demand for ‘high-level’ routine tasks like monitoring processes. These task shifts mainly occur within detailed occupations and become stronger over time. While displacement effects are...	Christina Gathmann, Felix Grimm, Erwin Winkler
9	2021	AI, firms and wages: Evidence from India ↗ This paper directly addresses the project's core questions by empirically estimating the causal effects of AI adoption on employment (vacancies) and wages using a novel dataset and exogenous variation in AI capability. It provides critical evidence on distributional outcomes, identifying specific winners and losers across skill levels and roles, which aligns with the project's focus on labor market reshaping and inequality. We examine the impact of artificial intelligence (AI) on hiring and wages in service sector firms, using a novel dataset of vacancy posts from India’s largest jobs website. We first document a rapid rise in demand for machine learning (ML) skills since 2016, particularly in the IT, finance and professional services industries. Vacancies requiring ML skills list substantially higher wages, but require more education and are highly concentrated both geographically and in the largest firms. Exploiting plausibly exogenous variation in exposure to advances in AI capabilities, we then examine the impacts of establishment demand for ML skills, as a proxy for AI adoption. We find that growth in the...	Katherine Stapleton, Alex Copestake, Ashley Pople
9	2025	Artificial intelligence, hiring and employment: job postings evidence from Sweden ↗ This paper directly addresses the project's core question of whether AI augments or substitutes for labor, providing empirical evidence from Swedish job postings that AI exposure increases hiring for both AI and non-AI roles. It offers crucial insights into firm-level labor reorganization and the distributional effects of AI adoption, aligning closely with the study of worker productivity and employment dynamics. This paper investigates the impact of artificial intelligence (AI) on hiring and employment, using the universe of job postings published by the Swedish Public Employment Service from 2014-2022 and universal register data for Sweden. We construct a detailed measure of AI exposure according to occupational content and find that establishments exposed to AI are more likely to hire AI workers. Survey data further indicate that AI exposure aligns with greater use of AI services. Importantly, rather than displacing non-AI workers, AI exposure is positively associated with increased hiring for both AI and non-AI roles. In the absence of substantial productivity gains that might account for this...	Erik Engberg, Mark Hellsten, Farrukh Javed, et al.
9	2023	From Occupations to Tasks: A New Perspective on Automatability Prediction Using BERT ↗ This paper directly addresses the core theme of measuring AI exposure at the task level rather than the occupation level, aligning with the project's task-based framework. It provides a novel methodological contribution using LLMs to predict automatability, which is central to understanding how AI reshapes labor markets and identifies winners and losers. As automation technologies continue to advance at an unprecedented rate, concerns about job displacement and the future of work have become increasingly prevalent. While existing research has primarily focused on the potential impact of automation at the occupation level, there has been a lack of investigation into the automatability of individual tasks. This paper addresses this gap by proposing a BERT-based classifier to predict the automatability of tasks in the forthcoming decade at a granular level leveraging the context and semantics information of tasks. We leverage three public datasets: O*NET Task Statements, ESCO Skills, and Australian Labour Market Insights Tasks, and perform...	Dawei Xu, Haoran Yang, Marian-Andrei Rizoiu, et al.
9	2024	The relationship between Artificial Intelligence (AI) exposure and returns to education ↗ This paper directly addresses the project's core question regarding the distributional effects of AI on wages by analyzing wage premiums across different skill levels and occupations in Europe. It provides critical empirical evidence on AI skill complementarity and inequality, which are central themes for understanding who the winners and losers are in the AI-driven labor market. Abstract This paper studies the relationship between exposure to artificial intelligence (AI) and workers’ wages across European countries. Overall, a positive relationship between exposure to AI and workers’ wages is found, however it differs considerably between workers and countries. High-skilled workers experience far higher wage premiums related to AI-related skills than middle- and low-skilled workers. Positive associations are concentrated among occupations moderately and highly exposed to AI (between the 6 th and 9 th decile of the exposure), and are weaker among the least exposed occupations. Returns of AI-related skills among high-skilled workers are even higher in Eastern...	Karol Madoń
9	2025	Generative AI and Labor Market Outcomes: Evidence from the United Kingdom ↗ [Title only] This title directly addresses the project's core interest in the causal effects of generative AI on labor market outcomes, providing relevant empirical evidence from a major economy. It aligns with the themes of aggregate labor market effects and distributional impacts, although the specific focus on the UK rather than the US may limit immediate generalizability to the researcher's primary context. No abstract available.	Bouke Klein Teeselink
9	2022	Ai, Skill, and Productivity: The Case of Taxi Drivers ↗ This paper directly addresses the project's core theme of AI skill complementarity versus substitution by providing empirical evidence that AI acts as a substitute for low-skilled labor. It specifically analyzes how AI reshapes productivity and inequality across worker skill levels, which is central to the researcher's investigation of who the winners and losers are in the AI labor market. We examine the impact of artificial intelligence (AI) on productivity in the context of taxi drivers. The AI we study assists drivers with finding customers by suggesting routes along which the demand is predicted to be high. We find that AI improves drivers’ productivity by shortening the cruising time, and this gain is accrued only to low-skilled drivers, narrowing the productivity gap between high- and low-skilled drivers by 13.4%. This case study provides evidence that AI and skill are indeed substitutes, offering direct support for the underlying assumption of recent projection exercises regarding job displacement by AI. This paper was accepted by Joshua Gans, business strategy...	Kyogo Kanazawa, Daiji Kawaguchi, Hitoshi Shigeoka, et al.
9	2022	What is the Price of a Skill? Revealing the Complementary Value of Skills ↗ This paper directly addresses the project's theme of AI skill complementarity versus substitution by empirically identifying AI as a 'hub skill' that enhances the value of other competencies through combination. It provides critical evidence on how AI reshapes labor market premiums and skill demand, informing the distributional effects and task reorganization aspects of the research. The global workforce is urged to constantly reskill, as technological change favours particular new skills while making others redundant. But which skills are most marketable and have a sustainable demand? We propose a model for skill evaluation that attaches a premium to a skill based on near real-time online labour market data. The model allows us to isolate the economic return of an individual skill measured as a premium on hourly wages. We demonstrate that the value of a specific skill is strongly determined by complementarity - that is with how many other high-value skills a competency can be combined. We introduce the idea of “hub skills” to the field of human capital formation, i.e...	Fabian Stephany, Ole Teutloff, Vili Lehdonvirta
9	2025	Are artificial intelligence skills a reward or a gamble? Deconstructing the AI wage premium in Europe ↗ This paper directly addresses the project's core theme of distributional effects by empirically identifying an AI wage premium for technical workers in Europe. It provides causal insights into how AI skills act as a reward, offering critical evidence on AI labor market inequality and task-specific wage dynamics. Understanding the labour market impact of new, autonomous digital technologies, particularly generative or other forms of artificial intelligence (AI), is currently at the top of the research and policy agenda. Many initial studies, though not all, have shown that there is a wage premium to mostly technical AI skills in labour markets. Such evidence tends to draw on data from web-based sources and typically fails to provide insight into the mechanisms underlying the AI wage gap. This paper utilises representative adult workforce data from 29 European countries, the second European skills and jobs survey, to examine wage differentials of the AI programmer workforce. The latter is uniquely...	Konstantinos Pouliakas, Giulia Santangelo, Paul Dupire
9	2025	AI and Immunity to the COVID-19 Pandemic ↗ This paper directly addresses the project's core question of whether AI tools augment or substitute for labor by providing empirical evidence that AI enables labor complementarity during economic disruptions. It also contributes to understanding the aggregate and sectoral effects of AI adoption on economic resilience and productivity in the context of a major external shock. We evaluate economic immunity to the COVID-19 pandemic accorded by a sector’s exposure to artificial intelligence (AI). Using an event study design, we causally estimate the impact of AI industry exposure (AIIE) on shifts in economic activity following the first pandemic-induced lockdown in India. We use payments data from one of India’s largest payment gateways and find that a one-standard-deviation increase in AIIE arrests the post-pandemic decline in industry payments by a third. High AIIE of neighboring sectors further improves a sector’s economic resilience, emphasizing the importance of supply chain partners in technology adoption and change. Further analyses speak to the mechanisms...	Abhishek Bhatia, Nikhil Madan, Deepa Mani, et al.
9	2024	Tasks, wages and new technologies ↗ This paper directly addresses the project's core theme of measuring AI exposure and its impact on wages within a task-based framework. It provides empirical evidence on how different types of technology, including AI, differentially affect wages for nonroutine versus routine tasks, which is central to understanding skill complementarity and distributional effects. Abstract This paper addresses the role of technology in shaping worker‐level task prices, exploiting within‐occupation variation using a unique survey linked to administrative data for over 180,000 Dutch workers between 2014 and 2020. Nonroutine abstract and interactive tasks are related to wage premia, and routine tasks to wage penalties. However, these task returns vary according to exposure to the types of (new) technology, such as computers, robots and artificial intelligence. Overall, wages are higher in technology‐intensive industries, but newer technologies target non‐routine tasks differently. This may have profound implications for the nonroutine wage premium given the rise of...	Femke Cnossen
9	2025	Artificial Intelligence and Labor Market Transformations in Latin America ↗ This paper directly addresses the project's core questions regarding AI exposure measurement, distributional effects on wages and inequality, and the distinction between augmentation and substitution across different worker groups. It provides relevant empirical evidence on how AI impacts labor markets, specifically highlighting differences by skill level and occupation, which aligns closely with the project's focus on winners and losers and aggregate labor market effects. This study examines the implications of artificial intelligence (AI) on employment, wages, and inequality in Latin America and the Caribbean (LAC). The paper identifies tasks and occupations most exposed to AI using comprehensive individual-level data alongside AI exposure indices. Unlike traditional automation, AI exposure correlates positively with higher education levels, ICT, and STEM skills. Notably, younger workers and women with high-level ICT and managerial skills face increased AI exposure, underscoring unique opportunities. A comparison of LAC with the OECD countries reveals greater impacts of AI in the former, with physical and customer-facing tasks showing divergent correlations...	Pablo Egaña-delSol, Claudio Bravo‐Ortega
9	2025	Displacement or Augmentation? The Effects of AI Innovation on Workforce Dynamics and Firm Value ↗ [Title only] The title directly addresses the core theme of whether AI augments or substitutes for labor, which is central to the project's investigation into workforce dynamics. It also aligns with the project's interest in how firms reorganize work and the resulting economic effects on firm value. No abstract available.	Mark A. Chen, Xiaoyu Wang
9	2025	Beyond Automation: Redesigning Jobs with LLMs to Enhance Productivity ↗ This paper directly addresses the project's core question of whether AI augments or substitutes for labor by empirically investigating job redesign through LLMs in the UK Civil Service. It provides detailed evidence on task-level AI exposure, productivity gains, and the reorganization of work, which are central themes in the researcher's project. The adoption of generative artificial intelligence (AI) is predicted to lead to fundamental shifts in the labour market, resulting in displacement or augmentation of AI-exposed roles. To investigate the impact of AI across a large organisation, we assessed AI exposure at the task level within roles at the UK Civil Service (UKCS). Using a novel dataset of UKCS job adverts, covering 193,497 vacancies over 6 years, our large language model (LLM)-driven analysis estimated AI exposure scores of 1,542,411 tasks. By aggregating AI exposure scores for tasks within each role, we calculated the mean and variance of job-level exposure to AI, highlighting the heterogeneous impacts of AI, even for...	Andrew Ledingham, Michael Hollins, Matthew Lyon, et al.
9	2025	AI and the Extended Workday: Productivity, Contracting Efficiency, and Distribution of Rents ↗ [Title only] This title directly addresses the core themes of AI's impact on worker productivity and the distribution of economic rents, which relates to wages and inequality. It also connects to firm reorganization and contracting efficiency, suggesting insights into how AI reshapes labor market structures and work organization. No abstract available.	Wei Jiang, Jun‐Young Park, Rachel J. Xiao, et al.
9	2024	Automation and Rent Dissipation: Implications for Wages, Inequality, and Productivity ↗ This paper directly addresses the project's core themes of automation's impact on wages, inequality, and productivity within a task-based framework. It provides key empirical evidence on how automation affects wage dispersion and offsets productivity gains, which is central to understanding the distributional and aggregate effects of technological change in labor markets. This paper studies the effects of automation in a task-based economy in which some jobs pay workers rents—wages above their outside option. We show that automation targets high-rent tasks, dissipating rents, amplifying wage losses, and reducing within-group wage dispersion in exposed groups. This form of rent dissipation is inefficient and offsets the productivity gains from automation. Using US data from 1980 to 2016, we find evidence of sizable rent dissipation and reduced within-group wage dispersion due to automation. Automation accounts for 52% of the increase in between-group inequality since 1980, with rent dissipation explaining one-fifth of this total. Our estimates imply that...	Daron Acemoğlu, Pascual Restrepo
9	2025	Organizational Transmission of AI: Managerial Influence on Generative AI Adoption ↗ This paper directly addresses the project's theme of how firms reorganize work and the factors driving AI adoption, specifically highlighting the crucial role of managerial influence and organizational context. It provides empirical evidence on the determinants of generative AI uptake and links these adoption patterns to employee engagement, which is central to understanding the distributional and productivity effects of AI in the labor market. Using longitudinal data from the Gallup Panel with roughly 10,000 U.S. respondents, surveyed annually between 2023 and 2025, we document new patterns in the heterogeneous adoption of generative AI and its "organizational transmission" within firms. We show that trust in leadership and clear managerial communication are the strongest predictors of employee uptake, even after accounting for income, occupation, and sector. Exploiting within-person variation and managerial exposure, we demonstrate that the complementarity between AI adoption and workplace culture significantly shapes employee outcomes. Specifically, employees who adopt AI in environments characterized by high managerial trust...	Christos Makridis
9	2025	Assessing the impact of AI skills on firm productivity: a cross-country analysis ↗ This paper directly addresses the core question of how AI-related skills affect firm productivity, a central theme of the research project. It provides empirical evidence on the causal link between AI skill availability and economic outcomes, which is essential for understanding the distributional and aggregate effects of AI adoption. Purpose Artificial intelligence (AI) is acknowledged for its long-term impact as a general-purpose technology, but it is also reshaping the skills and competences required in the job market. As discrepancies in workforce skills between countries play an important role in shaping differences in AI absorption, this paper explores the impact of evolving AI skills on firm productivity at the macro level, using a panel of 15 countries over the period 2017–2022. Design/methodology/approach The study employs the Relative AI Hiring Index from the Stanford Institute for Human-Centered AI to track changes in the AI-skilled labor force. A two-stage procedure is used to evaluate its impact on...	Ahmed Bounfour, Alberto Nonnis, Sae Jun Yang
9	2025	Generative AI Adoption in Human Creative Tasks: Experimental Evidence ↗ [Title only] This paper directly addresses the core themes of generative AI productivity experiments and task-based frameworks by providing experimental evidence on creative human tasks. It is highly relevant for understanding how AI tools augment or substitute labor and the resulting implications for worker productivity and inequality. No abstract available.	Wenbo Zou, Feng Zhu
9	2025	AI and the Future of Work in an Aging Economy ↗ [Title only] This title directly addresses the intersection of AI and labor markets, specifically focusing on the distributional effects across age groups, a core question in the researcher's project. The mention of an 'aging economy' suggests a unique perspective on how AI adoption might mitigate or exacerbate demographic shifts, aligning with themes of inequality and demographic-specific workforce impacts. No abstract available.	Carlo Pizzinelli, Marina Mendes Tavares
9	2025	AI Adoption and Inequality ↗ This paper directly addresses the project's core themes of AI's distributional effects on wages and inequality, as well as the mechanisms of labor substitution versus complementarity. It provides a calibrated task-based model analysis that specifically examines how AI adoption impacts different income groups and wealth disparities, aligning closely with the researcher's interest in winners and losers across skill levels. There are competing narratives about artificial intelligence’s impact on inequality. Some argue AI will exacerbate economic disparities, while others suggest it could reduce inequality by primarily disrupting high-income jobs. Using household microdata and a calibrated task-based model, we show these narratives reflect different channels through which AI affects the economy. Unlike previous waves of automation that increased both wage and wealth inequality, AI could reduce wage inequality through the displacement of high-income workers. However, two factors may counter this effect: these workers’ tasks appear highly complementary with AI, potentially increasing their productivity, and they...	Emma Rockall, Marina Mendes Tavares, Carlo Pizzinelli
9	2025	Tracking Employment Changes in AI-Exposed Jobs ↗ [Title only] This title directly addresses the core theme of AI exposure measurement and its empirical application to labor markets, which is central to the project's scope. It likely provides essential data or methodology for tracking how AI adoption affects different occupations, making it highly relevant for understanding distributional effects. No abstract available.	B Chandar
9	2023	A New Index Measuring Occupational Exposure to Artificial Intelligence ↗ This paper directly addresses the project's core question of how to measure AI exposure across occupations by introducing a novel semantic similarity index. It also critically examines the mechanism of whether AI substitutes or augments labor, specifically testing the 'Turing Transformation' hypothesis which posits that AI may broaden job opportunities rather than simply replacing workers. A central concern regarding artificial intelligence (AI) is its potential to replace human workers. However, recent research suggests that AI may improve the workers' job prospects through a Turing Transformation process: AI simplifies work, lowers barriers to job entry, and thus broadens job opportunities for more workers. In this paper, we present a conceptual framework that considers the two dimensions of AI's impact on job content and job opportunity , placing the Turing Transformation as a special case. We develop a novel occupational AI exposure measure using a sentence transformer model to compare the semantic similarity between the occupation descriptions and AI patents. We find...	Yuanyang Liu, Chuanren Liu, Tingliang Huang
9	2023	“This Time It's Different” Generative Artificial Intelligence and Occupational Choice ↗ This paper directly addresses the core question of how AI affects entry-level workers and the traditional career ladder by analyzing changes in apprenticeship search behavior following the launch of generative AI. It provides empirical evidence on occupational choice and automation risk perception, which are key mechanisms for understanding the distributional effects and labor market reshaping caused by AI. In this paper, we show the causal influence of the launch of generative AI in the form of ChatGPT on the search behavior of young people for apprenticeship vacancies. There is a strong and long-lasting decline in the intensity of searches for vacancies, which suggests great uncertainty among the affected cohort. Analyses based on the classification of occupations according to tasks, type of cognitive requirements, and the expected risk of automation to date show significant differences in the extent to which specific occupations are affected. Occupations with a high proportion of cognitive tasks, with high demands on language skills, and those whose automation risk had previously been...	Daniel Göller, Christian Gschwendt, Stefan C. Wolter
9	2025	AI occupational exposure and wage distribution: the case of Italy ↗ This paper directly addresses the project's core question regarding the distributional effects of AI on wages by analyzing how AI exposure impacts different parts of the wage distribution. It provides relevant empirical evidence on whether AI acts as a complement or substitute, specifically highlighting its role in exacerbating wage inequality among high earners. Questo articolo esamina la relazione tra l’esposizione all’intelligenza artificiale (IA) delle occupazioni e i salari guardando all’intera distribuzione salariale. A partire da un dataset di tipo employer-employee arricchito dalle informazioni derivanti da una survey contenente informazioni dettagliate sulle caratteristiche di ogni singola occupazione, stimiamo un modello di regressione quantilica non condizionata per un campione di lavoratori dipendenti italiani nel periodo 2011-2019. Al fine di verificare l’esistenza di un differenziale salariale attribuibile esclusivamente al tipo di occupazione, definito sulla base di un indicatore di esposizione potenziale che cattura soprattutto la...	Irene Brunetti, Chiara Mussida
9	2025	Implications of Artificial Intelligence and Robots for Employment and Labor Productivity: Firm-Level Evidence from the Republic of Korea ↗ This paper directly addresses the core questions regarding the causal effects of AI on labor productivity, employment, and the labor share of income. It provides firm-level evidence that distinguishes AI from robots, offering crucial insights into how AI adoption reshapes work organization and distributional outcomes. Examining data from firms in the Republic of Korea, this paper finds that artificial intelligence (AI) and robots differ in their impacts on employment and labor productivity. It finds that AI has a more positive overall impact on labor market outcomes. While both adopting AI and adopting robots increase employment, only adopting AI improves labor productivity. However, those productivity gains are associated with a decrease in the labor share of income. In addition, there is no evidence that firms adopting both robots and AI improve their labor productivity, potentially reflecting a lack of synergy.	Donghyun Park, Kwanho Shin
9	2026	Skill-biased technological change in the age of AI: a theoretical analysis of automation and inequality ↗ [Title only] This paper directly addresses the core theme of AI skill complementarity versus substitution and its distributional effects on inequality. It aligns well with the project's interest in how AI exposure varies by skill level and its macroeconomic implications for labor markets. No abstract available.	Te Li, Hadrian Nichols
9	2025	Shifting Work Patterns with Generative AI ↗ This paper directly addresses the project's core theme of generative AI productivity experiments by providing rigorous causal evidence from a large-scale field experiment on worker productivity and task time. It specifically investigates how AI adoption affects individual work patterns, such as time spent on email and working outside regular hours, which is central to understanding the labor market impacts of AI. We present evidence from a field experiment across 66 firms and 7,137 knowledge workers. Workers were randomly selected to access a generative AI tool integrated into applications they already used at work for email, meetings, and writing. In the second half of the 6-month experiment, the 80% of treated workers who used this tool spent two fewer hours on email each week and reduced their time working outside of regular hours. Apart from these individual time savings, we do not detect shifts in the quantity or composition of workers'tasks resulting from individual-level AI provision.	Eleanor Wiske Dillon, Sonia Jaffe, Nicole Immorlica, et al.
9	2026	The labor market effect of generative artificial intelligence on artists ↗ This paper directly addresses the project's core questions by empirically examining the causal effects of generative AI on employment, wages, and task reorganization within a specific creative labor market. It contributes key insights into AI skill complementarity versus substitution and the distributional impacts on workers, specifically focusing on the mechanisms of task reallocation and early-stage labor market adjustment. Technological change has repeatedly disrupted creative labor markets, raising concerns about whether new tools substitute for artists or shift the organization of creative work. This paper studies how occupational exposure to generative AI (genAI) maps into employment and earnings outcomes for U.S. artists following the unanticipated release of ChatGPT. I combine an occupation-level LLM task exposure index with establishment-based occupational outcomes from the occupational employment and wage statistics and individual microdata from the American Community Survey, estimating event-study specifications that compare more versus less exposed artistic occupations from 2017 to 2024. Across...	Christos Makridis
9	2025	The Labor Market Incidence of New Technologies ↗ This paper directly addresses the project's core theme of AI's distributional effects on wages and employment by developing a new theoretical framework for labor market incidence. It explicitly analyzes how AI and automation impact workers across skill dimensions, providing critical insights into the mechanisms of labor market adjustment and inequality relevant to the research goals. This paper develops a new framework to analyze the incidence of labor market shocks, focusing on automation and artificial intelligence. Central to our theory is the distance-dependent elasticity of substitution (DIDES), where worker mobility between occupations declines with their distance in skill space. Mapping 306 occupations into cognitive, manual, and interpersonal skill dimensions, we estimate a low-dimensional latent skill model that preserves granular substitution patterns. We show that both automation and artificial intelligence cluster within skill-adjacent occupations, constraining employment adjustment and amplifying wage effects. The clustering nature of technologies generates...	Tianyu Fan
9	2024	Artificial Intelligence and the Skill Premium ↗ This paper directly addresses the project's core question regarding who the winners and losers of AI are by analyzing its differential impact on the skill premium. It provides a theoretical framework to determine whether AI acts as a substitute or complement for high-skill workers, which is central to understanding distributional effects and AI skill complementarity versus substitution. What will likely be the effect of the emergence of ChatGPT and other forms of artificial intelligence (AI) on the skill premium? To address this question, we develop a nested constant elasticity of substitution production function that distinguishes between industrial robots and AI. Industrial robots predominantly substitute for low-skill workers, whereas AI mainly helps to perform the tasks of high-skill workers. We show that AI reduces the skill premium as long as it is more substitutable for high-skill workers than low-skill workers are for high-skill workers.	David E. Bloom, Klaus Prettner, Jamel Saadaoui, et al.
9	2025	Automation and Polarization ↗ This paper directly addresses the project's core themes by providing a theoretical framework for how automation reshapes labor markets, specifically focusing on wage polarization and the substitution of tasks based on complexity. It elucidates the mechanisms of skill complementarity versus substitution, detailing how changes in capital costs affect employment and wages across different skill levels, which is central to understanding the distributional effects of AI. We develop an assignment model of automation. Each of a continuum of tasks of variable complexity is assigned to either capital or one of a continuum of labor skills. We characterize conditions for interior automation, whereby tasks of intermediate complexity are assigned to capital. Interior automation arises when the most skilled workers have a comparative advantage in the most complex tasks relative to capital, and because the wages of the least skilled workers are suﬃciently low relative to their productivity and the eﬀective cost of capital in low-complexity tasks. Minimum wages and other sources of higher wages at the bottom make interior automation less likely. Starting with interior...	Daron Acemoğlu, Jonas Loebbing
9	2024	Generative AI & TeamWork: An experimental approach. ↗ [Title only] This paper directly addresses the core theme of generative AI productivity experiments by investigating how AI tools affect team dynamics and performance. It provides insight into whether AI augments or substitutes for labor within collaborative settings, a key question for understanding task reorganization and workforce productivity. No abstract available.	Fabrizio Dell’Acqua, Charles Ayoubi, Karim R. Lakhani
9	2025	The Economic Impact of Artificial Intelligence on Global Productivity and Labour Markets: A 2025 Perspective ↗ This paper directly addresses the project's core questions regarding the aggregate economic impact of AI on labor markets, including global productivity, employment patterns, and income distribution. It provides high-level evidence on job displacement versus creation and exposure rates, which are central to understanding the causal effects and distributional consequences of AI adoption. This research paper examines the transformative economic impact of artificial intelligence (AI) on global productivity and labour markets from a 2025 perspective. Drawing on extensive empirical evidence and economic projections, the study analyzes AI as an emerging General Purpose Technology (GPT) with profound implications for economic growth, employment patterns, and income distribution. The research reveals that while AI presents unprecedented opportunities for productivity gains, potentially adding $2.6-4.4 trillion annually to the global economy, it simultaneously poses significant challenges through job displacement and rising inequality. The findings indicate that 40% of global jobs...	Syed Ali Husain Rizvi
9	2025	Generative AI and Entrepreneurial Entry ↗ This paper directly addresses the project's core question on whether AI augments or substitutes for labor by analyzing how generative AI drives entrepreneurial entry through distinct augmentation and automation channels. It provides critical empirical evidence on distributional effects by skill level and occupation (STEM vs. non-STEM), offering insights into how firms and workers reorganize work in response to AI adoption. This study examines whether access to generative AI (GenAI) technologies affects entrepreneurial entry and, if so, how. We propose two mechanisms for a potential positive effect: (1) an augmentation channel that pulls prospective entrepreneurs into opportunity-driven entrepreneurship as they automate various peripheral tasks, and (2) an automation channel that pushes displaced wage workers into necessity-driven entrepreneurship as firms automate their core tasks. Leveraging the sudden release of ChatGPT, which democratized public GenAI access, we exploit industry variation in GenAI exposure for the science, technology, engineering, and mathematics (STEM) workforce in a...	Jiayi Bao, Bowen Lou, Hongshen Sun
9	2025	The Skill Premium Across Countries in the Era of Industrial Robots and Generative AI ↗ This paper directly addresses the project's core themes of wage inequality, skill premiums, and the differential impacts of AI versus other automation technologies. It provides macroeconomic evidence on whether AI acts as a complement or substitute for labor, which is central to understanding distributional effects and labor market reshaping. How do new technologies affect economic growth and the skill premium? To answer this question, we analyze the impact of industrial robots and artificial intelligence (AI) on the wage differential between low-skill and high-skill workers across 52 countries using counterfactual simulations. In so doing, we extend the nested CES production function framework of Bloom et al. (2025) to account for crosscountry income heterogeneity. Confirming prior findings, we show that the use of industrial robots tends to increase wage inequality, while the use of AI tends to reduce it. Our contribution lies in documenting substantial heterogeneity across income groups: the inequality-increasing effect of...	Marcos Júnio Ribeiro, Klaus Prettner
9	2025	AI Agents and Higher-Order Work ↗ This paper directly addresses the core question of whether AI tools augment or substitute for labor by demonstrating that AI agents shift effort from implementation to supervision, thereby complementing expert workers. It provides crucial empirical evidence on task reorganization and the distributional effects of AI, specifically highlighting how higher-order tasks and experienced workers benefit from this technological shift. <p>How do AI agents influence knowledge work? This paper finds that agents shift worker <span>effort from implementation to supervision, which especially benefits verifiable work </span><span>and expert workers. I analyze data from the coding platform Cursor to study agents </span><span>in software production. First, I find that workers have become less likely to produce </span><span>output manually after AI agents were introduced, and are more likely to delegate work </span><span>to agents. Second, workers use agents for abstract, higher-order tasks like delegation, </span><span>context gathering, and planning. Third, agents are used more frequently in settings </span><span>with...	Suproteem Sarkar
9	2025	How AI-Augmented Training Improves Worker Productivity ↗ [Title only] This title directly addresses the core theme of generative AI productivity experiments and their causal effects on worker productivity. It strongly aligns with the investigation of whether AI tools augment labor and how firms reorganize work to enhance output. No abstract available.	Didier Fouarge, Marie-Christine Fregin, Simon Janßen, et al.
9	2024	Artificial Intelligence, Tasks, Skills, and Wages: Worker-Level Evidence from Germany ↗ [Title only] This title directly aligns with the project's core themes of task-based frameworks, AI exposure measurement, and distributional effects on wages and skills. The focus on worker-level evidence provides empirical insight into who the winners and losers are, addressing key questions about wage inequality and skill complementarity versus substitution. No abstract available.	Michael Koch, Erik Engberg, Magnus Lodefalk, et al.
9	2024	Strategic Responses to Technological Change: Evidence from Online Labor Markets ↗ This paper directly addresses the project's core themes by analyzing how freelancers reorganize their work and adjust strategic responses to generative AI adoption in online labor markets. It provides empirical evidence on distributional effects, specifically highlighting how high-skill workers face greater adjustment costs and how workers shift targeting in response to AI-driven changes in labor demand and supply. In this project, we examine how freelancers changed their strategic positioning on an online work platform following the launch of ChatGPT in November 2022 - a major advance in AI technologies. We document that post-ChatGPT, freelancers bid on fewer jobs and reposition themselves by differentiating their distribution of bids (i.e., job applications) relative to their prior behavior. We disentangle heterogeneity in strategic responses by exploring how exposure to changes in demand or supply shape incumbent repositioning. We find that the launch of ChatGPT was associated with a short-term decrease in labor demand and an increase in labor supply, though these changes vary across work domains...	Shun Yiu, Robert Seamans, Manav Raj, et al.
9	2024	Generative AI and Firm-level Productivity: Evidence from Startup Funding Dynamics ↗ [Title only] This paper directly addresses the core theme of generative AI productivity experiments by linking AI adoption to firm-level outcomes via startup funding dynamics. It provides valuable evidence on how firms reorganize work and allocate resources in response to AI, which informs broader questions about aggregate productivity and firm-level efficiency gains. No abstract available.	David Heller, Dominik Asam
9	2023	Task-Interdependencies between Generative Ai and Workers ↗ This paper directly addresses the project's core theme of task reorganization by formalizing the microeconomic foundations of worker-Generative AI interactions. It specifically investigates how task interdependencies and the quality of worker-AI collaboration drive productivity, aligning closely with the research question of whether AI tools augment or substitute for labor. Our paper formalizes a production function to give microeconomic foundations for the adoption of Generative AI (GAI) within workplaces. The production function accounts for task-interdependencies, the worker-GAI interaction and indistinguishability between human-created and AI-generated outputs. We show that workers and GAI represent two distinct but interdependent sides of the production, that jointly generate a network externality in learning that drives productivity. We find that in open learning organizations favoring the worker-GAI interaction, GAI should be matched to workers based on their ability to detect errors. We analyze configurations where the worker-GAI interaction is...	Emmanuelle Walkowiak
9	2023	Productivity and Wages: What Was the Productivity–Wage Link in the Digital Revolution of the Past, and What Might Occur in the AI Revolution of the Future? ↗ This paper directly addresses the project's core themes by analyzing the historical and prospective distributional effects of technological change on wages and productivity, specifically focusing on skill-biased impacts. It explicitly connects past digital revolutions to the current AI era, discussing who the winners and losers are and how wage inequality evolves, which is central to the researcher's inquiry. Abstract Wages have been spreading out across workers over time – or in other words, the 90th/50th wage ratio has risen over time. A key question is, has the productivity distribution also spread out across worker skill levels over time? Using our calculations of productivity by skill level for the United States, we show that the distributions of both wages and productivity have spread out over time, as the right tail lengthens for both. We add Organization for Economic Co-Operation and Development (OECD) countries, showing that the wage–productivity correlation exists, such that gains in aggregate productivity, or GDP per person, have resulted in higher wages for workers at the top and...	Edward P. Lazear, Kathryn Shaw, Grant Hayes, et al.
9	2025	The Chatgpt Effect: Assessing Demand Shifts for it Skills in Germany ↗ [Title only] This paper directly addresses the project's core interest in distributional effects by examining how generative AI shifts demand for specific technical skills within a major labor market. It provides evidence on whether AI acts as a substitute or complement for IT workers, a key question regarding skill complementarity and occupation-specific impacts. No abstract available.	Katrin Rickmeier, Mariya Afonina
9	2026	Firms' GitHub Copilot adoption and labor market outcomes for software engineers ↗ This paper directly addresses the project's core questions regarding how generative AI adoption affects labor market outcomes, specifically focusing on software engineers and entry-level workers. It provides empirical evidence on whether AI tools augment or substitute for labor and examines the impact on hiring and skill composition, aligning closely with the study of firm reorganization and distributional effects. Abstract Using LinkedIn and GitHub data, this paper examines how firms' adoption of GitHub Copilot (GHC), a generative AI coding assistant, relates to software engineer (SWE) skills and labor outcomes. GHC adoption is associated with around a 3%–5% higher monthly probability of hiring SWEs, driven by entry‐level hires. New hires exhibit around 5% more non‐programming skills, with no decrease in coding skills. These findings are consistent with, for SWEs, GAI's productivity impacts and creation of new tasks outweighing potential displacement effects from automation of some SWE tasks.	Matthew Baird, Mar Carpanelli, Brian Xu, et al.
9	2026	<scp>Artificial intelligence</scp> adoption and the demand for managerial expertise ↗ This paper directly addresses the core theme of how firms reorganize work in response to AI adoption, specifically focusing on the demand for managerial expertise. It provides empirical evidence on how AI reshapes occupational tasks, shifting demand away from routine administrative functions toward interpersonal and coordination skills. Abstract Research Summary This paper examines how firms' adoption of artificial intelligence (AI) relates to the demand for managers and managerial skills. Using a skills‐based measure of AI adoption derived from Lightcast job postings, we show that firms with greater AI adoption post more managerial vacancies and a higher share of such vacancies than less intensive adopters. These relationships are strongest in manufacturing and among firms with higher research & development intensity. Greater AI adoption is also associated with shifts in managerial skill requirements toward interpersonal and growth‐oriented skills, including stakeholder management, creativity, and sales management, and...	Liudmila Alekseeva, José Azar, Mireia Giné, et al.
9	2026	The Macroeconomic effects of generative AI ↗ This paper directly addresses the macroeconomic effects of generative AI on labor markets, a core question of the research project. It specifically analyzes how AI adoption impacts productivity and wage dynamics across different worker groups, aligning closely with the study's focus on distributional effects and aggregate economic impacts. Our purpose is to estimate the macroeconomic impact of generative Artificial Intelligence (gen-AI). A theoretical model, based on a two-level CES production function, is developed to consider different elasticities of substitution between capital and labor, but differentiating between worker groups. Gen-AI is modeled as a potential enhancer of productivity for the different labor groups. We estimate the model for 67 countries over period 2022–2025. Results suggest that gen-AI contributed to increasing the productivity of most workers, regardless of their education, contract type, full or partial work time, and vulnerability level. This can be explained as, contrary to prior advances in this...	Raúl Katz, Juan Jung
9	2024	Macroeconomic Productivity Effects of Artificial Intelligence ↗ This paper directly addresses the core question regarding economywide employment and wage effects, specifically focusing on the aggregate macroeconomic productivity impacts of AI. It synthesizes theoretical and empirical evidence to evaluate whether AI-driven productivity gains match historical benchmarks, which is central to understanding the broader labor market implications of the technology. Abstract Some observers expect that the current wave of new tools based on artificial intelligence (AI) models, such as the large language models, will have strong effects on labor productivity. I present definitions and classifications that help understanding AI as an economic input. I then review theoretical and empirical arguments about macroeconomic productivity effects of AI and conclude that research has so far found no indication that productivity effects of the diffusion of AI are likely to be higher than those associated with the internet boom around the year 2000. While considerable uncertainty around future effects remains, a recent review and calibration exercise by Acemoglu, D...	Marianne Saam
9	2025	Large Language Models, Small Labor Market Effects ↗ [Title only] This title directly addresses the project's core inquiry into the causal effects of AI on labor markets, specifically focusing on the aggregate or macroeconomic impacts of Large Language Models. The framing suggests an analysis of whether widespread AI adoption leads to significant displacement or augmentation, which is central to understanding economywide employment and wage effects. No abstract available.	Anders Humlum, Emilie Vestergaard
9	2026	<span>AI Exposure, Cognitive Task Intensity, and the Complementarity–Substitution Divide: Evidence from Occupational Data</span> ↗ This paper directly addresses the project's core themes by investigating how AI exposure varies across cognitive task intensity and its associated wage effects, specifically testing the complementarity versus substitution divide. It provides empirical evidence on who the winners and losers are within the occupational distribution, aligning closely with the researcher's interest in distributional effects and labor market reshaping. This paper examines how AI exposure is associated with occupational wages across the cognitive task intensity distribution. Using OEWS wage data and O*NET task descriptions, we construct TF–IDF measures of AI exposure and cognitive task intensity and document a turning point at the 79th percentile of cognitive intensity. Below this threshold, AI exposure is positively associated with wages; above it, the marginal association is negative — applying to roughly one in five occupations. These patterns are stable across SOC major-group fixed effects, winsorization, and a leave-one-out sensitivity analysis covering 54 lexicon variants, but are not detectable using the Felten et al. (2021) index...	Toshiaki Mitsudome, Jingxuan Zhao
9	2026	Skill Atrophy and AI Productivity Measurement ↗ This paper directly addresses the project's core question of how to measure AI exposure and its causal effects on productivity by identifying specific biases in AI impact estimation. It provides a theoretical framework for understanding skill atrophy and its implications for wage inequality, which are central themes in the research agenda. <div> How should we measure the productivity effects of AI? Standard estimates condition <span>on current skill, but when AI substitutes for the cognitive effort that builds skill, skill </span><span>itself becomes endogenous to past AI use. We formalize pedagogical quality, the fraction </span><span>of learning-by-doing that survives AI delegation, and show it governs two measurement </span><span>biases: state-path divergence inflates within-worker panel estimates; spillover bias </span><span>degrades control groups. The biases generate a “scissors” pattern: panel and RCT </span><span>estimates diverge over time. The same mechanism produces an ability reversal in </span><span>long-run...	Tommaso Bondi, Gentry Johnson
9	2026	Artificial Intelligence, Productivity, and the Workforce: Evidence from Corporate Executives ↗ This paper directly addresses the project's core questions regarding the causal effects of AI on worker productivity, employment, and wage inequality by providing empirical evidence from corporate executives. It specifically examines task-based labor market shifts, such as the decline of routine roles and increased demand for skilled technical positions, which aligns with the themes of AI skill complementarity vs. substitution and labor reallocation. We use novel data from a survey of nearly 750 corporate executives to study the effects of artificial intelligence (AI) on productivity and the workforce. We document substantial heterogeneity in AI adoption across firms, with more than half having already invested, though many smaller firms are only beginning to do so. Labor productivity gains are positive, vary across sectors, and are expected to strengthen in 2026, with the largest effects concentrated in high-skill services and finance. These gains are not primarily driven by firms' capital deepening but instead reflect increases in revenue-based total factor productivity, closely associated with innovation-and demand-oriented channels...	Salome Baslandze, Zachary Edwards, John R. Graham, et al.
9	2026	Low Barriers, High Stakes Formal and Informal Diffusion of AI in the Workplace ↗ This paper directly addresses core themes of AI exposure measurement and distributional effects by documenting skill-biased diffusion patterns in the German labor market. It provides critical empirical context for understanding how firm-led adoption influences worker productivity and inequality, aligning closely with the project's focus on task reorganization and aggregate productivity effects. Artificial intelligence (AI) is diffusing rapidly in the workplace, yet firm-level indicators and aggregate productivity measures struggle to capture its early impact. This paper examines how AI spreads through both employer- and employee-initiated adoption and what this implies for measurement and productivity. Using a representative survey of nearly 10,000 employees in Germany, we document a high extensive but low intensive margin of usage: while 64 percent of employees use AI tools, only 20 percent use them frequently. Diffusion is strongly skill-biased and depends less on establishment and regional characteristics. Employer-led adoption is associated with higher usage intensity...	Melanie Arntz, Myriam Baum, Eduard Brüll, et al.
9	2026	AI and the Economy: The Four Rates Which Define The Future ↗ This paper directly addresses the project's core question regarding aggregate labor market effects by estimating structural displacement and wage-productivity decoupling through an integrated macroeconomic framework. It provides specific empirical forecasts on employment impacts and fiscal policy implications, aligning closely with the project's focus on economywide employment and wage effects of AI. <div> <div> <div> <p><span>So far the macroeconomic literature on artificial intelligence has offered competing forecasts of productivity, employment, price, and AI-sector revenue effects, derived in isolation from structural assumptions rather than from current empirical evidence. This paper develops an integrated framework calibrated to observed data and uses it to produce coherent joint forecasts across all four dimensions. We construct a Four-Rate Framework decomposing the AI transition into Innovation, Deployment, Adoption, and Displacement rates, combined with Workflow Completeness (</span><span>C</span><span>∗</span><span>), a bottom-up sectoral automation ceiling derived from...	benjamin verschuere, Angus Cameron
9	2026	Cognitive Automation and Sustainable Development: Global Task‐Level Evidence on Large Language Models and Labor Markets ↗ This paper directly addresses the project's core question by measuring AI exposure at the task level and analyzing its distributional effects across occupations and economies. It specifically focuses on large language models and their impact on knowledge-intensive sectors, providing critical empirical evidence on how AI reshapes labor markets and worker roles. ABSTRACT The rise of large language models (LLMs) is transforming the trajectory of traditional rule‐based automation, while their global impact on the labor market remains largely unexplored. To assess this transition, we develop a novel bottom‐up framework linking detailed task data to occupational structures across a broad spectrum of economies. Our findings reveal that LLM‐driven automation disproportionately impacts roles centered on information processing, administration, and managerial coordination compared to those in physical or manual domains. At the sectoral level, this impact translates into higher exposure for knowledge‐intensive sectors like finance, education, and...	Yuanfan Li, Rongrong Li, Qiang Wang
9	2024	Artificial Intelligence and the Skill Premium ↗ This paper directly addresses the project's core question regarding who the winners and losers are in the AI era by modeling the differential impacts of AI on low-skill versus high-skill workers. It explicitly examines AI's role as a complement or substitute for specific skill levels, providing theoretical insights into distributional effects and the skill premium. What will likely be the effect of the emergence of ChatGPT and other forms of artificial intelligence (AI) on the skill premium? To address this question, we develop a nested constant elasticity of substitution production function that distinguishes between industrial robots and AI. Industrial robots predominantly substitute for low-skill workers, whereas AI mainly helps to perform the tasks of high-skill workers. We show that AI reduces the skill premium as long as it is more substitutable for high-skill workers than low-skill workers are for high-skill workers.	David E. Bloom, Klaus Prettner, Jamel Saadaoui, et al.
9	2025	Does Machine Learning Shift Job Requirements? Impacts on Entry-Level Opportunities ↗ [Title only] This title directly addresses two core themes of the project: the impact of machine learning on job task composition and the specific consequences for entry-level workers and career ladders. It suggests an empirical investigation into how AI reshapes skill demands, which is central to understanding distributional effects and labor market reorganization. No abstract available.	Yuanyang Liu, Chuanren Liu, Tingliang Huang
9	2025	Generative AI, Adoption, and the Structure of Tasks ↗ This paper directly addresses the project's core themes by analyzing how generative AI alters task structures, firm adoption incentives, and worker incentives. It explicitly explores the mechanisms of AI as an augmentation tool and its differential effects on skilled versus less-skilled workers, which is central to understanding distributional outcomes and task reorganization. Abstract This chapter deals with the intersection of economics and technology and how they may shape the decision to adopt generative artificial intelligence (genAI). From the firm perspective, we discuss characteristics that make genAI different from past technologies, such as reduced changeover costs and reduced sensitivity to variability. We argue that, compared with past automation, these characteristics may spur adoption in tasks that are less frequent within an organization or more complex. We explore incentives for workers to adopt genAI: if genAI tends to augment less-skilled workers, workers with greater ability to critically evaluate the outputs of genAI may not have an incentive...	Ramayya Krishnan, Laurence Ales, Christophe Combemale
9	2025	How Retrainable are AI-Exposed Workers? ↗ This paper directly addresses the project's core question regarding the distributional effects of AI by quantifying the retraining potential for workers in AI-exposed occupations. It provides critical empirical evidence on whether AI substitutes or complements labor in the context of workforce transition, a key theme for understanding winners, losers, and career ladder dynamics. We document the extent to which workers in AI-exposed occupations can successfully retrain for AI-intensive work. We assemble a new workforce development dataset spanning over 1.6 million job training participation spells from all U.S. Workforce Investment and Opportunity Act programs from 2012-2023 linked with occupational measures of AI exposure. Using earnings records observed before and after training, we compare high AI exposure trainees to a matched sample of similar workers who only received job search assistance. We find that AI-exposed workers have high earnings returns from training that are only 25 percent lower than the returns for low AI exposure workers. However, training...	Benjamin Hyman, Benjamin Lahey, Karen X. Ni, et al.
9	2025	A Progress Report on the Economics of Artificial Intelligence: Impacts, Challenges, and Future Directions ↗ This paper directly addresses the core questions of the project by examining how AI impacts labor markets, productivity, and inequality. It provides a comprehensive review of AI's role as a general-purpose technology, covering distributional effects and policy implications central to the researcher's agenda. The rapid advancement of artificial intelligence (AI) marks a transformative shift for economies and the field of economics. As a general-purpose technology, AI automates complex tasks, drives productivity, and fosters cross-sectoral innovation, positioning it as a central force in economic change. This paper addresses three key questions: What distinguishes AI as an economic driver compared to past technologies? How does AI influence productivity, labor markets, and inequality? What policy measures are essential to maximize AI’s benefits while mitigating its risks? The review highlights AI’s distinct role in reshaping economic models, its delayed productivity gains typical of the Solow...	Alexander Steinmetz
9	2025	Skill-biased Productivity Effects of AI Adoption: Evidence from a Financial Shared Service Center Reform ↗ [Title only] The title explicitly addresses skill-biased productivity effects and AI adoption, directly aligning with the core themes of AI skill complementarity vs. substitution and generative AI productivity experiments. By focusing on a specific firm-level reform, it likely provides causal evidence on how AI reshapes labor markets and worker productivity, which are central questions of the research project. No abstract available.	Qiang Wu, Peixiang He, Peng Zhou
9	2026	The Microeconomics of Artificial Intelligence, by JoshuaGans ( <scp>MIT</scp> Press, Cambridge, Massachusetts, 2025) 434 pp., 6 × 9 in, 39 figures ↗ This book directly addresses the core themes of the research project by examining the microeconomic mechanisms of AI, including labor market effects, firm adoption, and task reorganization. It serves as a comprehensive theoretical and empirical foundation for understanding how AI reshapes labor markets and worker outcomes. Data sharing is not applicable to this article as no datasets were generated or analysed during the current study.	Eduard Talamàs
9	2026	Generative AI and Occupational Entry Barriers: The Labor-Supply Channel of Technological Change ↗ This paper directly addresses the project's core questions regarding how generative AI reshapes labor markets, specifically focusing on entry-level barriers, wage inequality, and the interaction between productivity and labor supply. It aligns closely with the themes of AI exposure measurement, distributional effects, and the labor-supply channel of technological change. <div> <div> <div> How will generative AI (GenAI) affect wage inequality across occupations? Beyond standard labor-demand effects, GenAI can reshape the wage distribution by lowering the expertise required to enter occupations, thereby expanding the pool of potential workers. We develop a general-equilibrium model in which changes in occupational wages reflect the interaction between these labor supply shifts and productivity gains, generating a productivity--scarcity race. Using O*NET task data and large language model (LLM) ratings, we construct novel measures of GenAI-induced changes in occupational expertise requirements and the resulting Potential Supply Shifts (PSS), alongside a...	Seyed Mahdi Hosseini Maasoum, Guy Lichtinger
9	2026	Augmentation Versus Substitution: Asymmetric Effects of Generative AI on Firm Performance and Share Prices ↗ This paper directly addresses the core project question of whether AI augments or substitutes for labor by developing and testing a theoretical framework that distinguishes these two channels using firm-level data and LLM-based task scoring. It provides crucial empirical evidence on how AI exposure varies by task type (augmentation vs. substitution) and links these differential effects to changes in firm performance, productivity, and market valuation, aligning closely with the project's focus on causal effects, skill complementarity, and firm reorganization. How does generative AI affect firm value? We develop a theoretical framework that distinguishes two channels: augmentation, where AI enhances human productivity while preserving firm-specific knowledge, and substitution, where AI replaces human labor but may erode differentiation rents. The model predicts that augmentation unambiguously increases firm value, whereas substitution has ambiguous effects because cost savings can be offset by the standardization of previously scarce expertise. To test these predictions, we score 18,796 O*NET task descriptions using a large language model along both dimensions, aggregate the scores by occupation using importance weights, and map them to firms...	Toghrul Aghbabali, Kee H. Chung
9	2026	Automation, Augmentation, and the Productivity Translation Lag: An Economic Framework for Generative AI ↗ This paper directly addresses the project's core theme of whether AI augments or substitutes labor by developing a framework that distinguishes the economic mechanisms and productivity multipliers of automation versus augmentation. It explicitly tackles the aggregate productivity effects and firm-level reorganization questions central to the researcher's investigation into the causal impacts of AI on labor markets. Generative AI has expanded the technological frontier across a broad range of cognitive tasks, yet its effects on aggregate productivity remain elusive. This paper develops the Productivity Translation Framework (PTF), a conceptual architecture that explains this gap by tracing how AI capability is translated into economic outcomes through seven causally linked stages: frontier expansion and socio-economic adoption, the dual mechanisms of automation and augmentation, their asymmetric productivity multipliers, market elasticity as demand-side moderator, marginal cost structure as supply-side moderator, competitive normalization and market restructuring, and the augmentation-to-automation...	Esteve Almirall
9	2026	Limits to Human-AI Collaboration: AI as Skill Multiplier? ↗ This paper directly addresses the project's core question of whether AI tools augment or substitute for labor by modeling the conditions under which AI acts as a skill multiplier rather than a democratizing force. It provides theoretical insights into distributional effects and inequality, specifically highlighting how domain knowledge determines whether AI adoption enhances or reduces productivity across different skill levels. Does AI democratize expertise or amplify skill gaps? Using a Bayesian information acquisition framework, we show that when AI unreliability is high, there exists an expertise threshold below which AI adoption reduces productivity. Above this threshold, domain knowledge acts as a filter that allows users to extract value from flawed outputs at low cost, while novices face verification barriers that negate efficiency gains. This formalizes conditions under which AI amplifies rather than equalizes productivity differentials in highstakes professional domains.	Ravi Prakash Ranjan, Ravi Pandey
9	2026	Generation–Verification Asymmetry Inversion and Apprenticeship Pipeline Collapse Under AI Substitution ↗ This paper directly addresses the core project question regarding entry-level workers and the traditional career ladder by modeling the collapse of apprenticeship pipelines due to AI substitution. It provides a formal theoretical framework for how AI reorganizes work and affects the distribution of skills, specifically highlighting the disproportionate negative impact on junior cohorts in knowledge-intensive professions. <div> For most of economic history, generation has been costly and verification cheap; tacit-capital institutions were optimized for that asymmetry. AI inverts the cost structure: generation becomes near-free while verification remains a Polanyi residual that is tacit and labor-intensive. We develop a tacit-capital production model with CES output in generation and verification, AI substitution at rate theta, and verification capacity acquired only through accumulated generation experience under h(tau) = tau^a. Two phenomena follow. The asymmetry inversion is a smooth equilibrium transition at theta_inv where the binding factor switches; markets handle it. The apprenticeship pipeline...	Alex Chengyu Li
9	2025	How Retrainable Are AI-Exposed Workers? ↗ [Title only] This title directly addresses the core project question of who the winners and losers are by examining the adaptability of workers exposed to AI, a key dimension of distributional effects. It likely investigates retraining outcomes, which is critical for understanding how AI reshapes labor markets and career trajectories for specific demographic or skill-based groups. No abstract available.	Benjamin Hyman, Ben Lahey, Karen Ni, et al.
9	2026	The Diffusion of Artificial Intelligence Across Firms: Evidence from Europe ↗ This paper directly addresses the core theme of firm AI adoption by developing a novel, large-scale indicator to measure realized AI usage across European firms. It provides critical empirical evidence on how workforce skills act as complements to AI adoption, informing the project's investigation into skill complementarity, firm reorganization, and the distributional effects of AI diffusion. We develop a novel firm-level indicator of Artificial Intelligence adoption in Europe (MAP-AI) by extracting information on AI usage from more than three million firm websites from Belgium, France, Germany, and Luxembourg (2016–2024) using a Large Language Model. The indicator captures realized AI adoption as signaled on their website rather than potential exposure. Our method allows to detect not only whether firms adopt AI, but also their role in the AI ecosystem and the type of AI technology they employ. Validation against human-coded benchmarks and external referenecs confirms high accuracy and external validity. We find that the share of AI-active firms grew from 1% in 2016 to 12% in...	Julio Garbers, Terry Gregory
9	2026	Generative AI and Labor Demand: Evidence from Flemish Job Vacancies ↗ This paper directly addresses the project's core interest in how generative AI reshapes labor demand, specifically focusing on the impact on entry-level workers and experience tiers as outlined in the research goals. By providing empirical evidence that AI reduces hiring demand primarily for roles with no prior experience in highly exposed occupations, it offers critical insights into the distributional effects and task reorganization dynamics central to the researcher's framework. This paper examines how generative AI is reshaping the composition of labor demand. Using administrative vacancy data of online job postings in Flanders (Belgium) from 2021 to 2025, I compare occupations with different predetermined exposure to generative AI across three experience tiers: no experience, some experience, and experienced. Exploiting the public release of ChatGPT as a common timing shock in a difference-in-differences design, I find that effects are concentrated in the top exposure quartile and emerge gradually, becoming statistically significant only in the third year after the launch. Highly exposed roles that require no prior work experience experience the highest estimated...	Jeroen Mahieu
9	2026	Skills are Power: AI in a Labor Market with Asymmetric Skill Substitution ↗ This paper directly addresses the project's core question of whether AI substitutes for or augments labor by developing a theoretical framework focused on asymmetric skill substitution and its distributional effects on wages and employment. It explicitly models the mechanics of AI exposure across different skill levels, providing a critical theoretical lens for understanding who the winners and losers are in the AI-driven labor market. We develop a novel perspective on how AI impacts the labor market. Higher-skill workers may perform tasks that require only lower skills but not the other way around. This generates overskilling where the same task is performed by labor of different skill levels. Skills are the (only) source of market power. Competing in a larger segment of jobs, higher skills command higher wages while securing a greater distance from non-employment, which is then only relevant for people at the bottom of the skill spectrum. This skill-based story of market power also suggests that the allocation of any economic surplus brought by AI is skewed towards the top of the skill distribution. Specific predictions...	Cheng Wang, Jie Luo
9	2025	The LLM Productivity Cliff: Threshold Productivity and AI-Native Inequality ↗ This paper directly addresses the project's core question of whether AI tools augment or substitute for labor by identifying a 'productivity cliff' that creates distinct winners and losers based on AI-native skills. It provides critical empirical evidence on the distributional effects of AI, specifically regarding skill complementarity and the widening inequality between workers with and without architectural literacy. We define the LLM productivity cliff as a threshold phenomenon: users who adopt an engineering mindset for working with LLMs attain step-change productivity, while others experience modest gains or outright slowdowns. Synthesizing 2025 evidence across software development, customer support, and labor markets, we show that outcomes are diverse and discontinuous, with performance grouped above and below a capabilities threshold. We operationalize this threshold as architectural literacy, not as marginal prompt skill but as a qualitative shift toward decomposition, orchestration, and systematic validation. We identify boundary conditions that make cliffs more likely (task complexity...	Francesco Bisardi
9	2022	Measuring the Occupational Impact of AI: Tasks, Cognitive Abilities and AI Benchmarks (Extended Abstract)* ↗ This paper directly addresses the core theme of AI exposure measurement by proposing a novel framework that maps tasks to cognitive abilities and AI benchmarks. It provides essential methodological insights for quantifying how AI reshapes labor markets across different occupations, which is a central question of the researcher's project. We present a framework for analysing the impact of AI on occupations. This framework maps 59 generic tasks from different occupational datasets to 14 cognitive abilities and these to a comprehensive list of 328 AI benchmarks used to evaluate research intensity in AI. The use of cognitive abilities as an intermediate layer allows for an identification of potential AI exposure for tasks for which AI applications have not been explicitly programmed. We provide insights into the abilities through which AI is most likely to affect jobs, and we show how some of the abilities where AI research is currently very intense are linked to tasks with comparatively limited labour input in the labour...	Songül Tolan, Annarosa Pesole, Fernando Martínez‐Plumed, et al.
9	2025	Methods for Evaluating Engineering Occupational Exposure to Artificial Intelligence ↗ This paper directly addresses the core theme of AI exposure measurement by systematically reviewing methods for assessing occupational exposure to AI in engineering. It provides critical methodological insights into how to quantify AI's potential impact on specific workforces, which is fundamental to the project's inquiry into measuring exposure across occupations and tasks. As artificial intelligence (AI) technologies advance rapidly, their impact on the engineering profession is potentially imminent. This paper provides a systematic review and comprehensive comparison of existing methods to assess occupational exposure to AI applications. By evaluating 36 papers and reports, we categorize the methods into five themes: expert opinions, qualitative surveys, quantitative approaches, deep learning techniques including Natural Language Processing (NLP), and analyses utilizing previous models. The strengths and weaknesses of each method are examined based on key metrics, including analysis time, strength of data sources, reliability, and specificity at a task...	Tamara Kecman, Maxwell Fingold, Susan McCahan
9	2026	Konjunkturschlaglicht: Drei Jahre ChatGPT: Auswirkungen von LLMs auf den deutschen Arbeitsmarkt ↗ [Title only] This paper directly addresses the researcher's core questions regarding the aggregate labor market effects of Large Language Models, specifically focusing on the German context as a key economy. It likely provides empirical evidence on AI exposure and distributional impacts, which are central to understanding how generative AI reshapes occupations and workers. No abstract available.	Morten Grinna Normann
9	2026	AI Skills Improve Job Prospects: Causal Evidence from a Hiring Experiment This paper directly addresses the project's core interest in distributional effects by providing causal evidence on how AI skills influence hiring outcomes across different occupations and worker demographics. It specifically examines who benefits from AI adoption, such as older workers or those with lower education, which aligns with the inquiry into winners and losers in the AI-driven labor market. The growing adoption of artificial intelligence (AI) technologies has heightened interest in the labor market value of AI related skills, yet causal evidence on their role in hiring decisions remains scarce. This study examines whether AI skills serve as a positive hiring signal and whether they can offset conventional disadvantages such as older age or lower formal education. We conducted an experimental survey with 1,725 recruiters from the United Kingdom, the United States and Germany. Using a paired conjoint design, recruiters evaluated hypothetical candidates represented by synthetically designed resumes. Across three occupations of graphic design, office assistance, and software...	Fabian Stephany, Ole Teutloff, Angelo Leone
9	2026	AI Skills Improve Job Prospects: Causal Evidence from a Hiring Experiment ↗ This paper directly addresses the project's core questions regarding the distributional effects of AI on workers, specifically how AI skills influence hiring outcomes across different occupations and demographic groups. It provides causal evidence on AI skill complementarity and substitution, offering critical insights into who wins or loses in the labor market due to AI adoption. The growing adoption of artificial intelligence (AI) technologies has heightened interest in the labor market value of AI related skills, yet causal evidence on their role in hiring decisions remains scarce. This study examines whether AI skills serve as a positive hiring signal and whether they can offset conventional disadvantages such as older age or lower formal education. We conducted an experimental survey with 1,725 recruiters from the United Kingdom, the United States and Germany. Using a paired conjoint design, recruiters evaluated hypothetical candidates represented by synthetically designed resumes. Across three occupations of graphic design, office assistance, and software...	Fabian Stephany, Ole Teutloff, Angelo Leone
9	2024	Generative AI & TeamWork: An experimental approach. ↗ [Title only] The title explicitly mentions Generative AI and teamwork, directly addressing the project's interest in generative AI productivity experiments and how firms reorganize work. The experimental approach aligns well with the goal of identifying causal effects of AI on worker productivity and task-based outcomes. No abstract available.	Fabrizio Dell’Acqua, Charles Ayoubi, Karim R. Lakhani
9	2026	AI exposure and wage inequality in U.S. cultural labour markets ↗ This paper directly addresses the project's core questions by empirically measuring the causal effects of AI exposure on wages and inequality within a specific labor market. It evaluates alternative methods for quantifying AI exposure and examines whether AI acts as a substitute or complement to labor, providing key distributional insights relevant to the project's themes. This paper examines wage inequality in U.S. cultural labour markets in the context of recent AI diffusion. Using individual-level data from the American Community Survey (ACS) for 2015 and 2024, we compare workers in arts, entertainment, and recreation with observationally similar workers in the rest of the economy and analyse wage differentials associated with occupational AI exposure within the cultural sector. Our empirical strategy relies on nearest-neighbour matching and combines ACS microdata with two occupation-level measures of AI exposure: the ability-based index of Felten et al. (2018) and the patent-task exposure index of Webb (2019). We document three main findings. First...	Juan Prieto Rodríguez, Juan Gabriel Rodriguez, Raquel Sebastian
9	2026	New Technologies and the Labor Share. A Share-Altering Representation ↗ This paper directly addresses the project's core themes by modeling the macroeconomic and distributional effects of AI on the labor share and wage inequality through a task-based framework. It explicitly analyzes whether AI acts as a substitute for labor by replacing tasks, which aligns with the research question regarding labor augmentation versus substitution and its aggregate consequences. We propose a share-altering framework with constant-returns-to-scale Cobb-Douglas production functions which accommodates for several forms of technical progress (capital-skill complementarity, robotization, AI-type technological change). The model is consistent with stylized facts, such as increasing wage inequality driven by skill-biased technical change or robotization, and provides insights into the potentially nuanced consequences of AI adoption on the wage distribution. We also extend the model to consider the production of new inputs. We show that the introduction of new inputs, despite their positive impact on the labor demand, will reduce the labor share whenever they replace tasks...	Giuliano Curatola, Alberto Dalmazzo
9	2026	From Techne Equalisation to Phronesis Premium: Generative AI and Seniority Bias in Ageing Europe ↗ This paper directly addresses the project's inquiry into who the winners and losers of AI adoption are, specifically examining distributional effects by age and the impact on entry-level workers. It provides empirical evidence on how generative AI reshapes labor market dynamics through task-based mechanisms of skill equalization and the premium on experiential judgment. Contrary to the prevailing view that older workers are disadvantaged by or unable to keep pace with technological change, this paper argues that generative AI may advantage them at the expense of younger ones. The mechanism operates through two channels. First, generative AI equalises procedural competence (techne) across occupations and within the same occupation and task, enabling average and lower-performing workers to reach the quality of top performers, an executional prosthesis that recovers skills depreciated with age. Second, as execution becomes cheap and replicable, the scarce factor shifts to skills that AI cannot easily replicate yet, such as contextual judgment and responsible...	Dora Moscato
9	2026	AI Technology Shocks, Task-Based Production, and Short-Run Macroeconomic Dynamics ↗ This paper directly addresses the macroeconomics of AI and the task-based framework, explicitly modeling how AI shocks affect short-run dynamics, inflation, and the natural real rate. It provides critical theoretical mechanisms for understanding the distinction between AI as a substitute (automation) versus a complement (productivity), which is central to the project's investigation of labor market effects and firm reorganization. This paper develops a tractable model of the short-run macroeconomics of artificial intelligence in a task-based economy with capital. The same real-side structure is embedded in a flexible-price environment, which pins down the natural allocation and natural real rate, and in a sticky-price New Keynesian environment, which determines inflation dynamics and monetary policy. AI arrives through two channels: an automation-frontier shock that expands the range of tasks performed by capital, and a machine-productivity shock that raises capital efficiency in already automated tasks. Motivating this distinction, sectoral evidence shows that AI is associated with lower price pressure and stronger...	Vladimir Kühl Teles
9	2026	Artificial Intelligence Investments and Expertise Erosion ↗ This paper directly addresses the project's question regarding the impact of AI on entry-level workers and traditional career ladders by modeling how automation erodes expertise accumulation. It provides a theoretical mechanism for why AI adoption might negatively affect the supply of high-expertise labor, a key theme in understanding the distributional effects and long-term labor market restructuring caused by AI. We study the feedback between automation and expertise accumulation. Building on Autor and Thompson's (2025) "expertise framework", we develop an overlapping generations model with endogenous on-the-job learning. Automation reduces learning in entry-level tasks, slowing expertise acquisition and worsening the economy's expertise distribution. Yet, we show that pre-existing automation decreases firms' willingness to pay for marginal improvements in automation. Indeed, these improvements benefit occupations requiring high human expertise, while the supply of workers with high expertise declines with prior automation. This weakens the dynamic complementarity between automation and...	Pierre Chaigneau, Evan Jo
9	2026	BEYOND THE SKILL PREMIUM: AI AND THE VALUE OF CREATIVITY ↗ This paper directly addresses the project's core questions on how AI reshapes labor markets by empirically distinguishing between creative, algorithmic, and physical tasks to measure differential impacts on wages and inequality. It provides crucial evidence on the mechanism of skill complementarity versus substitution, specifically highlighting how AI unbundles the college degree and redistributes earnings within and between educational groups. This paper classifies 894 occupations by their creative, algorithmic, and physical task content using O*NET data and measures AI's differential impact across these dimensions using CPS earnings microdata for 619,200 workers. AI complements creative work but substitutes for algorithmic work. This complementarity is driven by workers without a college degree-AI provides algorithmic scaffolding that previously required formal education. Quantile regressions reveal opposite distributional patterns: AI compresses inequality among non-college workers by lifting the bottom, while widening inequality among college graduates by amplifying the top. These patterns are consistent with AI unbundling the...	Hernán Moscoso Boedo
9	2025	The Algorithmic Divide: How Generative AI Rewrites the Rules of Pay ↗ [Title only] The title explicitly links generative AI to compensation, directly addressing the core question of wage effects and inequality. It suggests a focus on how AI algorithms determine pay, which aligns with firm reorganization and the distributional impacts of AI on workers. No abstract available.	Weining Zhang
9	2026	Labor Futures Under Artificial Intelligence: Scenarios for the Philippine Economy ↗ This paper directly addresses the project's core questions by analyzing AI's causal effects on employment, wages, and inequality within a task-based framework for a specific emerging economy. It provides valuable empirical evidence on worker augmentation versus substitution, task reconfiguration, and the distributional impacts of AI exposure across different occupational sectors. Rapid advances in generative artificial intelligence (AI) are reshaping the nature of work worldwide, yet their labor market implications remain highly uncertain and context-dependent. This paper examines the likely impacts of AI on labor in the Philippines through a prognostic, scenario-based approach that moves beyond deterministic forecasts. Drawing on three complementary evidence bases—task-level evidence on what generative AI can already do in practice, occupational exposure and complementarity analysis using Philippine labor force data, and firm- and worker-level evidence on AI adoption—the study develops an integrated framework linking AI capabilities, occupational structure, and...	Dominic Vincent Ligot
9	2026	Writing Code vs. Shipping Code: Productivity Effects Across Generations of AI Coding Tools ↗ This paper directly addresses the project's core question of whether AI tools augment or substitute for labor by providing empirical evidence of strong complementarity between AI coding agents and human effort. It also contributes to understanding firm reorganization and productivity effects by demonstrating how task-level gains attenuate across the production hierarchy to final output. How do the productivity effects of AI evolve across successive generations of tools, and to what extent do task-level gains ultimately translate into final output? We study these questions in the context of software development, using data on more than 100,000 GitHub developers combined with their AI usage telemetry. In a matched event study design, we find that autocomplete, interactive coding agents, and autonomous coding agents each significantly increase coding activity (“commits”), with respective cumulative effects of 40%, 140%, and 180%. These gains, however, attenuate sharply across the production hierarchy: the 180% cumulative effect falls to 50% for the number of projects, and to...	Mert Demirer, Leon Musolff, Liyuan Yang
8	2011	Skills, Tasks and Technologies: Implications for Employment and Earnings ↗ This paper establishes the task-based framework that serves as the theoretical backbone for analyzing how AI exposure and adoption affect labor markets through substitution and complementarity effects. Its discussion of skill-biased technological change, job polarization, and the reorganization of work provides essential foundational context for the project's inquiry into AI's distributional and causal effects on employment and wages. A central organizing framework of the voluminous recent literature studying changes in the returns to skills and the evolution of earnings inequality is what we refer to as the canonical model, which elegantly and powerfully operationalizes the supply and demand for skills by assuming two distinct skill groups that perform two different and imperfectly substitutable tasks or produce two imperfectly substitutable goods. Technology is assumed to take a factor-augmenting form, which, by complementing either high or low skill workers, can generate skill biased demand shifts. In this paper, we argue that despite its notable successes, the canonical model is largely silent on a number of central...	Daron Acemoğlu, David Autor
8	2015	Why Are There Still So Many Jobs? The History and Future of Workplace Automation ↗ This paper directly addresses core themes of the project, including the balance between AI substitution and complementation, distributional effects via labor market polarization, and the historical context of automation's impact on employment. It provides a theoretical framework for understanding task reorganization and comparative advantage, which is highly relevant to the project's inquiry into how AI reshapes labor markets. In this essay, I begin by identifying the reasons that automation has not wiped out a majority of jobs over the decades and centuries. Automation does indeed substitute for labor—as it is typically intended to do. However, automation also complements labor, raises output in ways that leads to higher demand for labor, and interacts with adjustments in labor supply. Journalists and even expert commentators tend to overstate the extent of machine substitution for human labor and ignore the strong complementarities between automation and labor that increase productivity, raise earnings, and augment demand for labor. Changes in technology do alter the types of jobs available and what those jobs...	David Autor
8	2019	Robots and Jobs: Evidence from US Labor Markets ↗ This paper is highly relevant as it employs a task-based and exposure-based framework to quantify the causal effects of automation on employment and wages, directly addressing the project's core themes on labor market impacts. Its methodological approach and findings on displacement effects provide crucial context for understanding how advanced technologies, including AI, reshape labor markets. We study the effects of industrial robots on US labor markets. We show theoretically that robots may reduce employment and wages and that their local impacts can be estimated using variation in exposure to robots—defined from industry-level advances in robotics and local industry employment. We estimate robust negative effects of robots on employment and wages across commuting zones. We also show that areas most exposed to robots after 1990 do not exhibit any differential trends before then, and robots’ impact is distinct from other capital and technologies. One more robot per thousand workers reduces the employment-to-population ratio by 0.2 percentage points and wages by 0.42%.	Daron Acemoğlu, Pascual Restrepo
8	2017	What can machine learning do? Workforce implications ↗ This paper directly addresses the project's core theme of workforce implications from machine learning, focusing on how AI reshapes human roles. It aligns closely with the investigation into whether AI augments or substitutes for labor and the resulting distributional effects on workers. Profound change is coming, but roles for humans remain	Erik Brynjolfsson, Tom M. Mitchell
8	2018	What Can Machines Learn and What Does It Mean for Occupations and the Economy? ↗ This paper provides a foundational task-based framework for measuring AI exposure by developing a 'Suitability for Machine Learning' metric across occupations, which directly addresses the project's core theme of how to measure AI exposure. It offers critical context on which tasks are affected by ML and the necessity of job redesign, informing the discussion on whether AI augments or substitutes for labor. Advances in machine learning (ML) are poised to transform numerous occupations and industries. This raises the question of which tasks will be most affected by ML. We apply the rubric evaluating task potential for ML in Brynjolfsson and Mitchell (2017) to build measures of “Suitability for Machine Learning” (SML) and apply it to 18,156 tasks in O*NET. We find that (i) ML affects different occupations than earlier automation waves; (ii) most occupations include at least some SML tasks; (iii) few occupations are fully automatable using ML; and (iv) realizing the potential of ML usually requires redesign of job task content.	Erik Brynjolfsson, Tom M. Mitchell, Daniel Rock
8	2017	Revisiting the risk of automation ↗ This paper directly addresses the measurement of AI exposure and the distinction between occupations and tasks, a core theme of the project. It provides critical empirical evidence on how task-based frameworks refine estimates of automation risk compared to broad occupational metrics. In light of rapid advances in the fields of Artificial Intelligence (AI) and robotics, many scientists discuss the potentials of new technologies to substitute for human labor. Fueling the economic debate, various empirical assessments suggest that up to half of all jobs in western industrialized countries are at risk of automation in the next 10 to 20 years. This paper demonstrates that these scenarios are overestimating the share of automatable jobs by neglecting the substantial heterogeneity of tasks within occupations as well as the adaptability of jobs in the digital transformation. To demonstrate this, we use detailed task data and show that, when taking into accounting the spectrum...	Melanie Arntz, Terry Gregory, Ulrich Zierahn
8	2022	How Artificial Intelligence Technology Affects Productivity and Employment: Firm-level Evidence from Taiwan ↗ This paper directly addresses the project's core questions regarding the causal effects of AI on firm-level productivity and employment using firm-level evidence. It also contributes to the distributional effects theme by analyzing how AI adoption alters workforce composition and labor demand by skill level. The effects of the rapid development of artificial intelligence (AI), a general-purpose technology, on firm performance is an emerging and crucial issue. This study examines the impact of AI technology on firms’ productivity and employee profiles. We use the keyword-matching method to parse the text of Taiwan patent grants, and obtain matched firm-level data on AI innovations in Taiwan's electronics industry for the 2002–2018 period. Empirical estimations indicate that AI technology is positively associated with productivity and employment. Meanwhile, non-AI patents also generate pro-productivity and pro-employment effects with a magnitude similar to that of AI technology. Inventing AI...	Chih‐Hai Yang
8	2021	The Janus face of artificial intelligence feedback: Deployment versus disclosure effects on employee performance ↗ This paper directly addresses the project's theme of how firms reorganize work and the causal effects of AI on worker productivity through a field experiment on AI-generated feedback. It provides empirical evidence on the tension between AI augmentation and worker resistance, offering insights into distributional effects based on employee tenure. Abstract Companies are increasingly using artificial intelligence (AI) to provide performance feedback to employees, by tracking employee behavior at work, automating performance evaluations, and recommending job improvements. However, this application of AI has provoked much debate. On the one hand, powerful AI data analytics increase the quality of feedback, which may enhance employee productivity (“deployment effect”). On the other hand, employees may develop a negative perception of AI feedback once it is disclosed to them, thus harming their productivity (“disclosure effect”). We examine these two effects theoretically and test them empirically using data from a field experiment. We...	Siliang Tong, Nan Jia, Xueming Luo, et al.
8	2021	The Robot Revolution: Managerial and Employment Consequences for Firms ↗ This paper closely aligns with the project's themes of firm reorganization, changes in employment by skill level, and the impact of automation on managerial roles, which serve as a critical historical analogue for AI adoption. Although it focuses on robots rather than AI specifically, its empirical findings on organizational structure, span of control, and task reassignment provide highly relevant contextual evidence for understanding how firms adjust to new technologies. As a new general-purpose technology, robots have the potential to radically transform employment and organizations. In contrast to prior studies that predict dramatic employment declines, we find that investments in robotics are associated with increases in total firm employment but decreases in the total number of managers. Similarly, we find that robots are associated with an increase in the span of control for supervisors remaining within the organization. We also provide evidence that robot adoption is not motivated by the desire to reduce labor costs but is instead related to improving product and service quality. Our findings are consistent with the notion that robots reduce variance...	Jay Dixon, Bryan Hong, Lynn Wu
8	2014	Polanyi's Paradox and the Shape of Employment Growth ↗ This paper directly addresses the project's interest in distributional effects and task-based frameworks by linking labor market polarization to Polanyi's Paradox, a key theoretical underpinning for AI substitution and augmentation. It provides relevant context on how tacit versus explicit knowledge influences employment growth, which is central to understanding who the winners and losers are in the AI era. In 1966, the philosopher Michael Polanyi observed, "We can know more than we can tell... The skill of a driver cannot be replaced by a thorough schooling in the theory of the motorcar; the knowledge I have of my own body differs altogether from the knowledge of its physiology." Polanyi's observation largely predates the computer era, but the paradox he identified-that our tacit knowledge of how the world works often exceeds our explicit understanding-foretells much of the history of computerization over the past five decades. This paper offers a conceptual and empirical overview of this evolution. I begin by sketching the historical thinking about machine displacement of human labor, and...	David Autor
8	2018	Low-Skill and High-Skill Automation ↗ This paper provides a foundational theoretical framework for analyzing how automation affects different skill levels and wage inequality, which are central to the project's inquiry into AI's distributional effects. Its task-based approach directly informs the measurement and mechanistic understanding of how AI substitutes or complements labor across occupations. We present a task-based model in which high- and low-skill workers compete against machines in the production of tasks. Low-skill (high-skill) automation corresponds to tasks performed by low-skill (high-skill) labor being taken over by capital. Automation displaces the type of labor it directly affects, depressing its wage. Through ripple effects, automation also affects the real wage of other workers. Counteracting these forces, automation creates a positive productivity effect, pushing up the price of all factors. Because capital adjusts to keep the interest rate constant, the productivity effect dominates in the long run. Finally, low-skill (high-skill) automation increases (reduces)...	Daron Acemoğlu, Pascual Restrepo
8	2020	Machine learning and human capital complementarities: Experimental evidence on bias mitigation ↗ This paper directly addresses AI skill complementarity by demonstrating how human domain expertise mitigates machine learning biases in a knowledge-work context. It provides empirical evidence on the interaction between specific human capital traits and AI tools, which is central to understanding how AI reshapes labor market dynamics and productivity. Abstract Research Summary The use of machine learning (ML) for productivity in the knowledge economy requires considerations of important biases that may arise from ML predictions. We define a new source of bias related to incompleteness in real time inputs, which may result from strategic behavior by agents. We theorize that domain expertise of users can complement ML by mitigating this bias. Our observational and experimental analyses in the patent examination context support this conjecture. In the face of “input incompleteness,” we find ML is biased toward finding prior art textually similar to focal claims and domain expertise is needed to find the most relevant prior art. We also...	Prithwiraj Choudhury, Evan Starr, Rajshree Agarwal
8	2018	A Method to Link Advances in Artificial Intelligence to Occupational Abilities ↗ This paper is closely related as it directly addresses the core theme of measuring AI exposure by proposing a method to link AI advances to occupational abilities. It provides a foundational framework for quantifying how AI reshapes labor markets at the occupation level, which is essential for understanding the distributional effects and task-based impacts central to the researcher's project. Prior episodes of automation have led to economic growth and also to many changes in the workplace. We expect the same from artificial intelligence (AI). The link between AI and labor is complex, however. To assist researcher and policymakers, we provide a method that links advances in AI to occupational abilities, and then aggregates from these abilities to the occupation level. We demonstrate the method by estimating which occupational descriptions have changed the most due to advances in AI between 2010 and 2015, and check our estimates using the Bureau of Labor Statistics scheduled update to occupational descriptions in 2016.	Edward W. Felten, Manav Raj, Robert Seamans
8	2018	Artificial Intelligence, Automation and Work ↗ This paper provides a foundational task-based framework for analyzing the opposing displacement and productivity effects of AI on labor demand and wages, which is central to the project's core questions. It directly addresses the mechanisms of skill complementarity versus substitution and the aggregate distributional effects of automation on income shares, offering key theoretical context for the research. This paper summarize a framework for the study of the implications of automation and AI on the demand for labor, wages, and employment. Our task-based framework emphasizes the displacement effect that automation creates as machines and AI replace labor in tasks that it used to perform. This displacement effect tends to reduce the demand for labor and wages. But it is counteracted by a productivity effect, resulting from the cost savings generated by automation, which increase the demand for labor in non-automated tasks. The productivity effect is complemented by additional capital accumulation and the deepening of automation (improvements of existing machinery), both of which further...	Daron Acemoğlu, Pascual Restrepo
8	2021	Algorithm-Augmented Work and Domain Experience: The Countervailing Forces of Ability and Aversion ↗ This paper directly addresses the core theme of whether AI tools augment or substitute for labor by examining how domain experience influences worker performance when using algorithmic advice. It provides empirical evidence on the non-linear relationship between worker ability and AI adoption, highlighting specific mechanisms like algorithmic aversion that explain variation in AI effectiveness across different skill and experience levels. Past research offers mixed perspectives on whether domain experience helps or hurts algorithm-augmented worker performance. Reconciling these perspectives, we theorize that intermediate levels of domain experience are optimal for algorithm-augmented performance, due to the interplay between two countervailing forces—ability and aversion. Although domain experience can increase performance via increased ability to complement algorithmic advice (e.g., identifying inaccurate predictions), it can also decrease performance via increased aversion to accurate algorithmic advice. Because ability developed through learning by doing increases at a decreasing rate, and algorithmic aversion is more...	Ryan Allen, Prithwiraj Choudhury
8	2023	What is the price of a skill? The value of complementarity ↗ The paper directly investigates AI skill complementarity, a core mechanism for understanding how AI affects wages and labor market distribution. It provides empirical evidence on how AI skills augment worker productivity and value through combinations with other skills, addressing the project's interest in winners and losers by skill level. The global workforce is urged to constantly reskill, as technological change favours particular new skills while making others redundant. But which skills are a good investment for workers and firms? As skills are seldomly applied in isolation, we propose that complementarity strongly determines a skill's economic value. For 962 skills, we demonstrate that their value is determined by complementarity – that is, how many different skills, ideally of high value, a competency can be combined with. We show that the value of a skill is relative, as it depends on the skill background of the worker. For most skills, their value is highest when used in combination with skills of a different type...	Fabian Stephany, Ole Teutloff
8	2023	Economics of ChatGPT: a labor market view on the occupational impact of artificial intelligence ↗ This paper directly addresses the core theme of AI exposure measurement by providing a systematic classification of occupational susceptibility to generative AI. It aligns closely with the project's focus on how AI reshapes labor markets, although it relies on potential rather than realized causal effects. Purpose The study investigates the influence of ChatGPT on the labor market dynamics, aiming to provide a structured understanding of the changes induced by generative AI technologies. Design/methodology/approach An analysis of existing literature serves as the foundation for understanding the impact, while the supply and demand model helps assess the effects of ChatGPT. A text-mining approach is utilized to analyze the International Standard Occupation Classification, identifying occupations most susceptible to disruption by ChatGPT. Findings The study reveals that 32.8% of occupations could be fully impacted by ChatGPT, while 36.5% might experience a partial impact and 30.7% are likely to...	Ali Zarifhonarvar
8	2023	The Uneven Impact of Generative AI on Entrepreneurial Performance ↗ This paper directly addresses the distributional effects of AI by examining how generative AI impacts firm performance differently across skill levels, aligning with the project's focus on winners and losers. It provides empirical evidence on AI skill complementarity, showing that high-performing entrepreneurs benefit while low-performers decline, which is central to understanding inequality and task reorganization. Scalable and low-cost AI assistance has the potential to improve firm decision-making and economic performance. However, running a business involves a myriad of open-ended problems, making it difficult to know whether recent AI advances can help business owners make better decisions in real-world markets. In a field experiment with Kenyan entrepreneurs, we assessed the impact of AI advice on small business revenues and profits by randomizing access to a GPT-4-powered AI business assistant via WhatsApp. While we are unable to reject the null hypothesis that there is no average treatment effect, we find the treatment effect for entrepreneurs who were high performing at baseline to be 0.27...	Nicholas G. Otis, Rowan Philip Clarke, Solène Delecourt, et al.
8	2020	Digitalization, routineness and employment: An exploration on Italian task-based data ↗ This paper provides a task-based analysis of how digitalization affects employment, directly addressing the project's core theme of measuring AI exposure through routine task intensity and digital use. Its findings on routine-biased technological change offer valuable empirical context for understanding how automation technologies reshape labor markets and occupational structures. This paper explores the relation between the digitalization of labour processes, the level of routineness of labour tasks and changes in employment in the case of Italy in the period 2011-16. The levels of digitalization and routineness of occupations in more than 500 4-digit ISCO professional groups are measured using data from a unique Italian profession-level survey on skill, tasks and work contents – the INAPP-ISTAT Survey on Italian Occupations (ICP), an O*NET-type dataset. Two digitalization indices are used: a digital use index, measuring the use of computers and e-mail in the workplace, and a digital tasks index, capturing the presence of a set of key digital tasks, such as those...	Valeria Cirillo, Rinaldo Evangelista, Dario Guarascio, et al.
8	2021	New digital technologies and heterogeneous wage and employment dynamics in the United States: Evidence from individual-level data ↗ This paper directly addresses the project's core questions by empirically estimating the heterogeneous effects of AI and digital technologies on individual wages and employment dynamics using robust longitudinal data. It provides valuable evidence on whether these technologies act as substitutes or complements to labor and identifies specific demographic groups, such as highly educated workers, who are most impacted. We analyze heterogeneous effects of new digital technologies on individual-level wage and employment dynamics in the United States from 2011-2018. To this end, we employ four digital technology measures from recent literature: computerization probabilities of occupations, occupational impacts of artificial intelligence, and the suitability of tasks for machine learning and their within-occupation variance. Based on CPS and ASEC panel data, the results indicate that labor-displacing digital technologies are associated with slower wage growth and higher probabilities of switching one's occupation and becoming non-employed. In contrast, labor-reinstating digital technologies improve individual...	Frank M. Fossen, Alina Sorgner
8	2020	Robot Imports and Firm-Level Outcomes ↗ This paper closely relates to the project's core themes by examining the causal effects of automation technology on firm-level employment, productivity, and wage inequality, using a rigorous empirical strategy. While it focuses on industrial robots rather than AI specifically, its findings on task displacement and skill complementarity provide valuable mechanistic insights for understanding how new technologies reshape labor markets. We use French data over the 1994-2013 period to study how imports of industrial robots affect firm-level outcomes. Guided by a simple model, we develop a novel empirical strategy to identify the causal effects of robot adoption. Our results suggest that, while demand shocks generate a positive correlation between robot imports and employment at the firm level, exogenous exposure to automation leads to job losses. We also find that robot exposure increases labour productivity and some evidence that it may raise the relative demand for high-skill professions.	Alessandra Bonfiglioli, Rosario Crinò, Harald Fadinger, et al.
8	2024	The power of generative marketing: Can generative AI create superhuman visual marketing content? ↗ This paper provides empirical evidence on how generative AI augments specific creative tasks in visual marketing, directly addressing the project's theme of AI skill complementarity versus substitution. The findings on productivity gains and task reorganization offer valuable insights into the distributional effects of AI on specialized occupations and firm-level production processes. Generative AI’s capacity to create photorealistic images has the potential to augment human creativity and disrupt the economics of visual marketing content production. This research systematically compares the performance of AI-generated to human-made marketing images across important marketing dimensions. First, we prompt seven state-of-the-art generative text-to-image models (DALL-E 3, Midjourney v6, Firefly 2, Imagen 2, Imagine, Stable Diffusion XL Turbo, and Realistic Vision) to create 10 , 320 synthetic marketing images, using 2 , 400 real-world, human-made images as input. 254 , 400 human evaluations of these images show that AI-generated marketing imagery can surpass human-made...	Jochen Hartmann, Yannick Exner, Samuel Domdey
8	2024	From Man vs. Machine to Man + Machine: The art and AI of stock analyses ↗ This paper directly addresses the core theme of whether AI tools augment or substitute for labor by demonstrating human-AI synergy in complex financial tasks. It provides specific empirical evidence on how workers adapt to AI capabilities and the conditions under which human judgment adds incremental value, informing the project's investigation into task-based frameworks and labor market reorganization. An AI analyst trained to digest corporate disclosures, industry trends, and macroeconomic indicators surpasses most analysts in stock return predictions. Nevertheless, humans win “Man vs. Machine” when institutional knowledge is crucial, e.g., involving intangible assets and financial distress. AI wins when information is transparent but voluminous. Humans provide significant incremental value in “Man + Machine”, which also substantially reduces extreme errors. Analysts catch up with machines after “alternative data” become available if their employers build AI capabilities. Documented synergies between humans and machines inform how humans can leverage their advantage for better adaptation...	Sean Cao, Wei Jiang, Junbo L. Wang, et al.
8	2024	AI adoption in America: Who, what, and where ↗ This paper directly addresses the firm AI adoption theme by providing empirical evidence on the diffusion patterns and determinants of early AI technologies across U.S. firms. It offers valuable context for understanding the distributional aspects of the 'AI divide' and how adoption varies by firm size, location, and owner characteristics, which are key to assessing broader labor market impacts. Abstract We study the early adoption and diffusion of five artificial intelligence (AI)‐related technologies (automated‐guided vehicles, machine learning, machine vision, natural language processing, and voice recognition) as documented in the 2018 Annual Business Survey of 850,000 firms across the United States. We find that fewer than 6% of firms used any of the AI‐related technologies we measure, though most very large firms reported at least some AI use. Weighted by employment, average adoption was just over 18%. AI use in production, while varying considerably by industry, was found in every sector of the economy and clustered with emerging technologies, such as cloud computing and...	Kristina McElheran, J. Frank Li, Erik Brynjolfsson, et al.
8	2021	A review on the economics of artificial intelligence ↗ This paper provides a comprehensive macroeconomic review of AI's economic impacts, directly addressing core themes such as aggregate labor market effects, inequality, and the distinction between AI and previous technologies. It synthesizes empirical evidence on employment and skill-based disparities, offering valuable context for the project's investigation into how AI reshapes labor markets. Abstract The rapid development of artificial intelligence (AI) not only represents a scientific breakthrough but also has impacts on human society and economies, as well as on the development of economics. This paper focuses on the macroeconomic perspective, reviewing recent literature in order to answer three key questions. First, what approaches are being used to represent AI in economic models? Second, will AI technology have an impact on the economy different from that of previous new technologies? Third, in which aspects will AI have an impact, and what is the empirical evidence for these effects of AI? On the first question, our review reveals that the incorporation of AI into...	Yingying Lu, Yixiao Zhou
8	2021	Human–AI collaborative decision-making as an organization design problem ↗ This paper directly addresses the project's theme of how firms reorganize work and the division of labor between humans and AI. It provides a conceptual framework for understanding the mechanisms of augmentation versus substitution, which is central to the research questions regarding task reorganization and skill complementarity. Abstract The promise of collaboration between humans and algorithms in producing good decisions is stimulating much experimentation. Drawing on research in organization design can help us to approach this experimentation systematically. I propose typologies for considering different forms of division of labor between human and algorithm as well as the learning configurations they are arranged in, as basic building blocks for this endeavor.	Phanish Puranam
8	2023	Generative AI and jobs ↗ This paper directly addresses the project's core question on how AI exposure varies across occupations and whether AI acts as a substitute or complement to labor, specifically highlighting the augmenting nature of Generative AI. It provides crucial context on distributional effects, particularly regarding gender and income levels, which aligns with the project's focus on who the winners and losers are. This study assesses the potential global exposure of occupations to Generative AI, particularly GPT-4. It predicts that the overwhelming effect of the technology will be to augment occupations, rather than to automate them. The greatest impact is likely to be in high and upper-middle income countries due to a higher share of employment in clerical occupations. As clerical jobs are an important source of female employment, the effects are highly gendered. Insights from this study underline the need for proactive policies that focus on job quality, ensure fair transitions, and that are based on dialogue and adequate regulation.	Paweł Gmyrek, Janine Berg, David Bescond
8	2023	AI-Driven Productivity Gains: Artificial Intelligence and Firm Productivity ↗ This paper directly addresses the core theme of AI's causal effects on firm productivity, providing empirical evidence that aligns with the researcher's interest in productivity gains and technological upgrading. It also touches upon the mechanism of skill-biased enhancement, which is relevant to understanding how AI interacts with labor skills and firm organization. Artificial intelligence is profoundly influencing various facets of our lives, indicating its potential to significantly impact sustainability. Nevertheless, capturing the productivity gains stemming from artificial intelligence in macro-level data poses challenges, leading to the question of whether artificial intelligence is reminiscent of the “Solow paradox”. This study employs micro-level manufacturing data to investigate the impact of artificial intelligence on firms’ productivity. The study finds that every 1% increase in artificial intelligence penetration can lead to a 14.2% increase in total factor productivity. This conclusion remains robust even after conducting endogeneity...	Xueyuan Gao, Hua Feng
8	2022	Automation and the Workforce: A Firm-Level View from the 2019 Annual Business Survey ↗ This paper is closely related as it provides firm-level empirical evidence on AI adoption and its impact on labor productivity, skill requirements, and employment, directly addressing key themes of the project. It offers valuable context on how advanced technologies are reshaping workforces and contributes to the discussion on AI's distributional effects and task reorganization. This paper describes the adoption of automation technologies by US firms across all economic sectors by leveraging a new module introduced in the 2019 Annual Business Survey, conducted by the US Census Bureau in partnership with the National Center for Science and Engineering Statistics (NCSES). The module collects data from over 300, 000 firms on the use of five advanced technologies: AI, robotics, dedicated equipment, specialized software, and cloud computing. The adoption of these technologies remains low (especially for AI and robotics), varies substantially across industries, and concentrates on large and young firms. However, because larger firms are much more likely to adopt them...	Daron Acemoğlu, Gary L. Anderson, D.K. Beede, et al.
8	2023	Transforming boundaries: how does ChatGPT change knowledge work? ↗ This paper directly addresses the core theme of whether AI augments or substitutes labor by developing a framework for generative AI's interaction with knowledge workers based on task type. It provides relevant qualitative insights into firm-level work reorganization and the changing nature of knowledge work, which are central to understanding the causal effects of AI adoption. Purpose This paper aims to demonstrate how the new generative artificial intelligence (AI) tool ChatGPT changes knowledge work for individuals and what are the implications of this change for companies. Design/methodology/approach Based on 22 interviews from informants across different industries, the authors conducted an inductive analysis on the use and utility of ChatGPT in knowledge work. Based on this initial analysis, they discovered different ways in which ChatGPT either augments human agency, makes it redundant or lacks capability in that regard. Findings The authors develop a 2 × 2 framework of algorithmic assistance, which demonstrates four ways in which ChatGPT (and generative AI...	Paavo Ritala, Mika Ruokonen, Laavanya Ramaul
8	2020	A Strategic Framework for Task Automation in Professional Services ↗ This paper directly addresses the core theme of task-based frameworks and the distinction between augmentation and substitution, specifically within professional services. It provides relevant theoretical context for understanding how firms reorganize work and how AI exposure varies across tasks within high-skill occupations. Professional service jobs exist at the high end of the skill ladder; thus, some have assumed that highly trained professional workers are relatively immune to being replaced by automation. However, this assumption is a bit dubious because automation does not occur at the job level but rather at the task level, and some tasks within a professional job might be highly susceptible to automation disruption. This research builds on prior research by (1) empirically testing a model for automation of professional services and (2) developing a professional task-automation framework that shows how individual tasks within a given job can be enhanced or disrupted by automation in very different ways...	Scott E. Sampson
8	2022	For whom the bell tolls: The firm-level effects of automation on wage and gender inequality ↗ This paper directly addresses the project's focus on distributional effects and wage inequality by examining how AI adoption impacts within-firm wage disparities and gender gaps. It provides relevant empirical evidence on whether AI tools augment or substitute for labor through its findings on wage increases and hiring patterns in adopting firms. This paper investigates the impact of investment in automation- and AI-related goods on within-firm wage inequality in the French economy during the 2002–2017 period. We document that most wage inequality in France is accounted for by differences among workers belonging to the same firm rather than by differences between sectors, firms, and occupations. Using an event-study approach on a sample of firms importing automation- and AI-related goods, we find that spike events related to the adoption of automation- and AI-related capital goods are not followed by an increase in within-firm wage inequality or in gender wage inequality. Instead, wages increase by 1% three years after the events at...	Giacomo Domini, Marco Grazzi, Daniele Moschella, et al.
8	2024	Digital transformation and labor upgrading ↗ This paper directly addresses the project's theme of labor market reshaping by examining how digital transformation drives demand for higher-skilled labor while reducing demand for production workers. It provides empirical evidence on skill complementarity and firm-level reorganization, offering valuable context for understanding the distributional effects and task-based shifts associated with technological adoption. Prior literature shows that digitalization can reduce transaction costs and improve operating efficiency. However, the success of enterprise digitalization heavily depends on a variety of complementary human capital and employee capabilities. We focus on whether and to what extent enterprise digitalization affects employee upgrading, and whether digitalization is more valuable in firms where digitalization matches with human capital. Our analysis measures employee upgrading using detailed employee-level data (i.e., experience and education) and matches these data to metrics on enterprise digitalization to determine whether a firm's digitalization facilitates employee upgrading. Using the...	Wencong Li, Xingquan Yang, Xingqiang Yin
8	2023	AI Adoption in America: Who, What, and Where ↗ This paper directly addresses the core theme of firm AI adoption, providing empirical evidence on which firms, industries, and geographic locations are early adopters. It offers crucial context for understanding the distributional effects and potential 'AI divide' by highlighting correlations with firm size, owner characteristics, and innovation strategies, which informs the broader questions about winners and losers in the labor market. We study the early adoption and diffusion of five artificial intelligence (AI)‐related technologies (automated‐guided vehicles, machine learning, machine vision, natural language processing, and voice recognition) as documented in the 2018 Annual Business Survey of 850,000 firms across the United States. We find that fewer than 6% of firms used any of the AI‐related technologies we measure, though most very large firms reported at least some AI use. Weighted by employment, average adoption was just over 18%. AI use in production, while varying considerably by industry, was found in every sector of the economy and clustered with emerging technologies, such as cloud computing and robotics...	Kristina McElheran, J. Frank Li, Erik Brynjolfsson, et al.
8	2013	The 'Task Approach' to Labor Markets: An Overview ↗ [Title only] This paper provides a theoretical foundation for the task-based framework, which is explicitly listed as a core theme for measuring AI exposure and understanding augmentation versus substitution. While it is an overview rather than an empirical study of AI itself, it is highly relevant for establishing the methodological basis needed to analyze how AI reshapes labor markets across occupations and tasks. No abstract available.	David Autor
8	1999	Multi-Task Learning and the Reorganization of Work. From Tayloristic to Holistic Organization This paper directly addresses the project's core theme of how firms reorganize work, specifically analyzing the shift from specialized to holistic task structures. It provides a theoretical framework for understanding technological complementarity and task reorganization, which are central to assessing AI's impact on labor markets. The paper analyzes the contemporary organizational restructuring of production and work within firms. We emphasize the shift from a "Tayloristic" organization of work (characterized by significant specialization by tasks) to a "holistic" organization (featuring job rotation, integration of tasks and learning across tasks). We examine four driving forces behind this restructuring process: advances in production technologies promoting technological task complementarities, advances in information technologies promoting informational task complementarities, changes in worker preferences in favor of versatile work, and advances in human capital that make workers more versatile. Our analysis can...	Assar Lindbeck, Dennis J. Snower
8	2021	Automation in Latin America: Are Women at Higher Risk of Losing Their Jobs? ↗ This paper directly addresses the project's core theme of distributional effects by measuring AI automation risk through a task-based framework while highlighting significant gender disparities. It provides crucial empirical context on how AI exposure varies across demographic groups, which is central to identifying winners and losers in the labor market. The Fourth Industrial Revolution, which comprises digitization, artificial intelligence, robotics, among others, have the power to drastically increase economic output but may also displace workers. In this paper we assess the risk of automation for female and male workers in four Latin American countries – Bolivia, Chile, Colombia and El Salvador. Our study is the first to apply a task-based approach with a gender perspective in this region. Our main findings indicate that men are more likely than women to perform tasks linked to the ‘skills of the future’, such as STEM (science, technology, engineering and mathematics), information and communications technology, management and...	Pablo Egaña-delSol, Monserrat Bustelo, Laura Ripani, et al.
8	2024	Does artificial intelligence promote common prosperity within enterprises? —Evidence from Chinese-listed companies in the service industry ↗ This paper directly addresses the project's core question regarding the distributional effects of AI on wages and inequality by analyzing the impact of AI on labor income share in the service sector. It provides relevant empirical evidence on how AI adoption affects different worker groups, specifically highlighting the displacement of low-educated and frontline workers, which aligns with the themes of AI substitution and labor market polarization. As the largest industry that absorbs labor from different levels of employment and provides different levels of labor remuneration, the service industry has faced severe challenges from the wave of artificial intelligence replacement. This study examines whether artificial intelligence promotes shared prosperity among service industry enterprises based on microdata of listed companies in the Chinese service industry from 2008 to 2022. The main research results indicate that the application of artificial intelligence in the service industry significantly reduces enterprises' labor income share. The main mechanisms of action include the employment structure effect of squeezing out...	Kaiming Chen, Xiaoqian Chen, Zhan-ao Wang, et al.
8	2024	The consequences of generative AI for online knowledge communities ↗ This paper directly addresses the project's interest in entry-level workers and task reorganization by documenting how generative AI adoption leads to the exit of junior users from online knowledge platforms. It provides relevant empirical evidence on the substitution effects of AI on specific tasks and the distributional impact on less experienced workers. Generative artificial intelligence technologies, especially large language models (LLMs) like ChatGPT, are revolutionizing information acquisition and content production across a variety of domains. These technologies have a significant potential to impact participation and content production in online knowledge communities. We provide initial evidence of this, analyzing data from Stack Overflow and Reddit developer communities between October 2021 and March 2023, documenting ChatGPT's influence on user activity in the former. We observe significant declines in both website visits and question volumes at Stack Overflow, particularly around topics where ChatGPT excels. By contrast, activity...	Gordon Burtch, Dokyun Lee, Zhichen Chen
8	2021	The impact of artificial intelligence on skills at work in Denmark ↗ This paper directly addresses the project's core themes of AI skill complementarity versus substitution and the distributional effects of AI on labor markets. It provides empirical evidence on how AI adoption reshapes task requirements and work organization, highlighting varied impacts across skill levels and occupations. Abstract Based on a unique dataset on the use of artificial intelligence (AI) among employees in Denmark, we investigate within‐job relationships between AI use and skill requirements. We show that the effects of AI are varied and depend on whether AI is used for providing orders to humans or providing information for further human handling and in which occupation it is used. AI may enhance or augment skills through, for example, the increased use of high‐performance work practices, or it may increase work pace constraints and reduce employee autonomy. The results imply that the diffusion of AI can increase inequalities in the labour market by augmenting skills used in high‐skill jobs...	Jacob Rubæk Holm, Edward Lorenz
8	2023	Influence of artificial intelligence applications on total factor productivity of enterprises—evidence from textual analysis of annual reports of Chinese-listed companies ↗ This paper directly addresses the project's core theme of aggregate labor market effects by empirically linking AI adoption to enterprise total factor productivity. It provides relevant evidence on whether AI acts as a substitute for labor, noting that positive TFP effects are driven by the replacement of low-end labor, which informs the discussion on who the winners and losers are. Artificial intelligence (AI) empowers the real economy, promotes intelligent transformation, and upgrades enterprises. However, whether AI applications improve enterprises’ total factor productivity (TFP) in developing countries remains unknown. Based on a textual analysis of the annual reports of Chinese A-share listed companies, we constructed indicators to measure AI applications in companies. Furthermore, the development status and influencing factors of AI applications in Chinese enterprises were explored, and the influence of AI applications on TFP was examined. The results reveal that the probability of AI application varies across enterprises. Large enterprises with a low proportion...	Yilin Zhong, Feng Xu, Longpeng Zhang
8	2018	Will This Time Be Different? A Review of the Literature on the Impact of Artificial Intelligence on Employment, Incomes and Growth ↗ This paper directly addresses the project's core question regarding whether AI differs from past technological shocks in its impact on labor markets and growth. It provides a critical review of the theoretical mechanisms, such as task-based models and complementarity versus substitution effects, which are central to understanding the causal effects of AI on wages and employment. There is a long-standing economic research literature on the impact of technological innovation and automation in general on employment and economic growth. Traditional economic models trade off a negative displacement or substitution effect against a positive complementarity effect on employment. Economic history since the industrial revolution as well as more recent evidence strongly supports the view that the net effect on employment and incomes is positive. Still, there are concerns that with AI "this time may be different". State-of-the-art economic models for predicting the impact of AI on occupations make it different by design, by emphasizing the labour substitution effects of this...	Bertin Martens, Songül Tolan
8	2023	Generative AI and Jobs: A Global Analysis of Potential Effects on Job Quantity and Quality ↗ This paper directly addresses the core theme of AI exposure measurement by providing a global assessment of occupational exposure to generative AI. It also engages with key distributional questions by analyzing the gendered impacts and the distinction between task augmentation versus automation. This study assesses the potential global exposure of occupations to Generative AI, particularly GPT-4. It predicts that the overwhelming effect of the technology will be to augment occupations, rather than to automate them. The greatest impact is likely to be in high and upper-middle income countries due to a higher share of employment in clerical occupations. As clerical jobs are an important source of female employment, the effects are highly gendered. Insights from this study underline the need for proactive policies that focus on job quality, ensure fair transitions, and that are based on dialogue and adequate regulation.	Paweł Gmyrek, Janine Berg, David Bescond
8	2022	Artificial Intelligence and Firm-Level Productivity ↗ This paper directly investigates the causal effect of AI adoption on firm-level productivity, addressing a core theme of the project regarding the economic impacts of AI. By using an IV approach to handle endogeneity, it provides robust empirical evidence on whether AI acts as an augmenting technology that enhances output, which is central to understanding labor market implications. Artificial Intelligence (AI) is often regarded as the next general-purpose technology with a rapid, penetrating, and far-reaching use over a broad number of industrial sectors. A main feature of new general-purpose technology is to enable new ways of production that may increase productivity. So far, however, only very few studies investigated likely productivity effects of AI at the firm-level; presumably because of lacking data. We exploit unique survey data on firms’ adoption of AI technology and estimate its productivity effects with a sample of German firms. We employ both a cross-sectional dataset and a panel database. To address the potential endogeneity of AI adoption, we also...	Dirk Czarnitzki, Gastón P. Fernández, Christian Rammer
8	2023	Artificial intelligence adoption and system‐wide change ↗ This paper directly addresses the project's core theme of how firms reorganize work in response to AI adoption by analyzing the systemic and interdependent nature of tasks within organizations. It provides a theoretical framework for understanding how AI integration requires broader structural changes beyond individual task substitution, which is critical for assessing aggregate labor market effects and skill complementarity. Abstract Analyses of artificial intelligence (AI) adoption focus on its adoption at the individual task level. What has received significantly less attention is how AI adoption is shaped by the fact that organizations are composed of many interacting tasks. AI adoption may, therefore, require system‐wide change, which is both a constraint and an opportunity. We provide the first formal analysis where multiple tasks may be part of an interdependent system. We find that reliance on AI, a prediction tool, increases decision variation, which, in turn, raises challenges if decisions across the organization interact. Reducing inter‐dependencies between decisions softens that impact and can...	Ajay Agrawal, Joshua S. Gans, Avi Goldfarb
8	2023	Examining the influence mechanism of artificial intelligence development on labor income share through numerical simulations ↗ This paper directly addresses the project's interest in distributional effects and the task-based framework by analyzing how AI-driven task substitution and creation influence labor income share. It provides relevant theoretical mechanisms regarding the balance between wage suppression and new task creation, which are central to understanding AI's impact on inequality and labor markets. Maintaining the stability of labor income share is a key foundation for optimizing the structure of income distribution. Based on Chinese macroeconomic data, this study constructs a dynamic general equilibrium model with production tasks and applies numerical simulation to investigate the impact of artificial intelligence (AI) development on labor income share. The results demonstrate that the influence direction of AI development on labor income share depends on the relative speed of machine replacement and new tasks. Mechanism analysis reveals that machine replacement makes the wage growth rate smaller than the labor productivity growth rate, causing a decline in labor income share...	Cheng Qian, Chun Zhu, Duen-Huang Huang, et al.
8	2023	The Impact of AI on the Workplace: Evidence from OECD Case Studies of AI Implementation ↗ This paper directly addresses the project's core themes of task reorganization and the augmentation versus substitution of labor by providing qualitative evidence that AI primarily reorients jobs rather than displacing workers. It offers valuable insights into firm-level responses and worker well-being, contributing to the understanding of distributional effects and the causal mechanisms of AI adoption in the workplace. How artificial intelligence (AI) will impact workplaces is a central question for the future of work, with potentially significant implications for jobs, productivity, and worker well-being. Yet, knowledge gaps remain in terms of how firms, workers, and worker representatives are adapting. This study addresses these gaps through a qualitative approach. It is based on nearly 100 case studies of the impacts of AI technologies on workplaces in the manufacturing and finance sectors of eight OECD countries. The study shows that, to date, job reorganisation appears more prevalent than job displacement, with automation prompting the reorientation of jobs towards tasks in which humans have a...	Anna Milanez
8	2024	The Rapid Adoption of Generative AI ↗ This paper directly addresses the core theme of firm and worker AI adoption by providing empirical evidence on the rapid uptake of generative AI in the US workforce. It contributes to understanding the aggregate effects and distributional patterns of AI exposure by quantifying usage rates and time savings across different industries and employment contexts. Generative artificial intelligence (genAI) is a potentially important new technology, but its impact on the economy depends on the speed and intensity of adoption. This paper reports results from a series of nationally representative U.S. surveys of genAI use at work and at home. As of late 2024, 45% of the U.S. population age 18–64 uses genAI. Among employed respondents, 27% used genAI for work at least once in the previous week: 10% used it every workday and 17% on some but not all workdays. Relative to each technology’s first mass-market product launch, work adoption of genAI has been faster than the personal computer (PC), and overall adoption has outpaced both PCs and the internet by...	Adam Blandin, Alexander Bick, David Deming
8	2022	Can digital skill protect against job displacement risk caused by artificial intelligence? Empirical evidence from 701 detailed occupations ↗ This paper directly addresses the project's core question of how AI reshapes labor markets by empirically measuring displacement risk and wage effects across detailed occupations. It specifically contributes to the themes of AI exposure measurement and distributional effects by demonstrating how digital skills moderate these risks, aligning with the inquiry into who are the winners and losers. To identify the role of digital skill in the skill-biased technological changes caused by artificial intelligence, this study estimates the impacts of displacement risk on occupational wage and employment and examines the moderation effects of digital skill through the occupational data from the U.S. Bureau of Labor Statistics through the methods of fixed-effects modeling, heterogeneity analyzing and moderation effect testing. The results highlight three main points that (1) the displacement risk by artificial intelligence has significantly negative effects on occupational wage and employment, (2) the heterogeneous effects across occupational characteristics are significant, and (3) the...	Ni Chen, Zhi Li, Bo Tang
8	2023	AI and the Accounting Profession: Views from Industry and Academia ↗ This paper directly addresses the project's core themes by examining how generative AI reshapes labor markets within a specific professional sector, focusing on task substitution and the displacement of entry-level workers. It provides empirical context on firm reorganization and skill complementarity, offering insights into the distributional effects of AI adoption on wages and employment. ABSTRACT Anecdotal and empirical evidence indicates that the growing adoption of artificial intelligence (AI) within accounting firms and accounting departments leads to improvements in efficiency, a gradual increase in the share of AI workers, and a decrease in junior accounting employees. If this trend continues, would it signal the beginning of an era of diminishing demand for new accounting professionals and a shift in the required skill set of new accounting employees? The aim of the workshop, which, by happenstance, occurred the same week that OpenAI introduced ChatGPT, was to bring together Accounting Information Systems researchers and representatives from leading accounting firms...	J. Efrim Boritz, Theophanis C. Stratopoulos
8	2024	The Economic Impacts and the Regulation of AI: A Review of the Academic Literature and Policy Actions ↗ This paper provides a comprehensive synthesis of economic research on AI's effects on employment, productivity, and inequality, directly addressing the project's core themes on aggregate labor market outcomes. It also evaluates the current state of empirical evidence and regulatory landscapes, offering valuable context for understanding the causal effects and distributional impacts of AI adoption. We review the literature on the effects of Artificial Intelligence (AI) adoption and the ongoing regulatory efforts concerning this technology. Economic research encompasses growth, employment, productivity, and income inequality effects, while regulation covers market competition, data privacy, copyright, national security, ethics concerns, and financial stability. We find that: (i) theoretical research agrees that AI will affect most occupations and transform growth, but empirical findings are inconclusive on employment and productivity effects; (ii) regulation has focused primarily on topics not explored by the academic literature; (iii) across countries, regulations differ widely in...	Mariarosaria Comunale
8	2024	Future jobs: analyzing the impact of artificial intelligence on employment and its mechanisms ↗ This paper closely addresses the project's core question regarding the causal effects of AI on aggregate employment, providing empirical evidence on whether AI substitutes or augments labor in the manufacturing sector. It directly contributes to the macroeconomics of AI theme by analyzing mechanisms like productivity gains and industry agglomeration, though it focuses on a specific national context rather than occupation-level task reorganization. Technological innovation has promoted the development of human flourishing. Based on panel data for 30 provinces in China from 2006 to 2022, this study examines the impact of artificial intelligence (AI) on manufacturing employment in China using the two-way fixed-effect model and the instrumental variable method. The study finds that contrary to the traditional impression of “machines replacing humans,” AI technology is correlated with increasing the total number of jobs on the market. Thanks to more efficient labor productivity, capital deepening, and specialized division of labor from integrating digital technology, AI offsets the negative effect of robots on employment and significantly...	Yang Shen
8	2021	From Man vs. Machine to Man + Machine: The Art and AI of Stock Analyses ↗ This paper directly addresses the core question of whether AI tools augment or substitute for labor, providing empirical evidence of a 'man plus machine' complementarity in high-skill financial analysis. It contributes to the project's themes of AI exposure measurement and distributional effects by showing how AI advantages vary by task complexity and information type. An AI analyst we build to digest corporate financial information, qualitative disclosure and macroeconomic indicators is able to beat the majority of human analysts in stock price forecasts and generate excess returns compared to following human analyst. In the contest of "man vs machine," the relative advantage of the AI Analyst is stronger when the firm is complex, and when information is high-dimensional, transparent and voluminous. Human analysts remain competitive when critical information requires institutional knowledge (such as the nature of intangible assets). The edge of the AI over human analysts declines over time when analysts gain access to alternative data and to in-house AI...	Sean Cao, Wei Jiang, Junbo Wang, et al.
8	2024	"If the Machine Is As Good As Me, Then What Use Am I?" – How the Use of ChatGPT Changes Young Professionals' Perception of Productivity and Accomplishment ↗ This paper directly addresses the project's core themes of AI augmentation, task-based frameworks, and the experiences of entry-level or young professionals. It provides qualitative insights into how generative AI reshapes worker perception and task delegation, offering valuable context on the psychological and productivity dimensions of AI adoption. Large language models (LLMs) like ChatGPT have been widely adopted in work contexts. We explore the impact of ChatGPT on young professionals’ perception of productivity and sense of accomplishment. We collected LLMs’ main use cases in knowledge work through a preliminary study, which served as the basis for a two-week diary study with 21 young professionals reflecting on their ChatGPT use. Findings indicate that ChatGPT enhanced some participants’ perceptions of productivity and accomplishment by enabling greater creative output and satisfaction from efficient tool utilization. Others experienced decreased perceived productivity and accomplishment, driven by a diminished sense of ownership...	Charlotte Kobiella, Y. López, Franz Waltenberger, et al.
8	2024	Automation and Augmentation: Artificial Intelligence, Robots, and Work ↗ This paper directly addresses the core themes of AI augmentation versus substitution and the distributional effects of automation on inequality and worker well-being. It provides relevant background on how specific job characteristics influence exposure to AI, aligning closely with the project's focus on task-based frameworks and labor market impacts. This article reviews the literature that examines the potential, limitations, and consequences of robots and artificial intelligence (AI) in automation and augmentation across various disciplines. It presents key observations and suggestions from the literature review. Firstly, displacement effects from task automation continue to persist. However, one should not assume an unequivocally increasing efficacy of technology in automation or augmentation, especially given the declining productivity growth in high-income countries and some large emerging economies in recent decades. Jobs less likely to be negatively impacted are those that require diverse tasks, physical dexterity, tacit...	Ya-Wen Lei, Rachel Kim
8	2023	Does Artificial Intelligence Promote or Inhibit On-the-Job Learning? Human Reactions to AI at Work ↗ This paper directly addresses the project's core theme of who are the winners and losers by analyzing AI's disproportionate negative impact on on-the-job learning for older, less-educated, and female workers. It also examines task reorganization and worker reorganization, specifically how AI adoption inhibits skill acquisition and motivation through mechanisms like burnout and reduced disposable time. This paper examines how AI at work impacts on-the-job learning, shedding light on workers’ reactions to the groundbreaking AI technology. Based on theoretical analysis, six hypotheses are proposed regarding three aspects of AI’s influence on on-the-job learning. Empirical results demonstrate that AI significantly inhibits people’s on-the-job learning and this conclusion holds true in a series of robustness and endogeneity checks. The impact mechanism is that AI makes workers more pessimistic about the future, leading to burnout and less motivation for on-the-job learning. In addition, AI’s replacement, mismatch, and deskilling effects decrease people’s income while extending working hours...	Chao Li, Yuhan Zhang, Xiaoru Niu, et al.
8	2025	Skills or degree? The rise of skill-based hiring for AI and green jobs ↗ This paper closely relates to the project by empirically analyzing how AI exposure reshapes hiring standards and wage structures, specifically highlighting the shift from formal degrees to specific technical skills in AI occupations. It provides direct evidence on the distributional effects of AI adoption, identifying winners and losers based on skill sets versus credentials, which addresses the core themes of AI skill complementarity and labor market inequality. Emerging professions in fields like Artificial Intelligence (AI) and sustainability (green jobs) are experiencing labour shortages as industry demand outpaces labour supply. In this context, our study aims to understand whether employers have begun focusing more on individual skills rather than formal qualifications in their recruitment processes. We analysed a large time-series dataset of approximately eleven million online job vacancies in the UK from 2018 to mid-2024, drawing on diverse literature on technological change and labour market signalling. Our findings provide evidence that employers have initiated “skill-based hiring” for AI roles, adopting more flexible hiring practices to...	Matthew Bone, Eugenia González Ehlinger, Fabian Stephany
8	2023	Investigating the influence of digital technology application on employee compensation ↗ This paper directly addresses the project's interest in AI/digital technology's effects on wages and inequality, specifically examining how technological adoption impacts average employee compensation and the executive-employee compensation gap. It provides relevant empirical evidence on the distributional consequences of technology, highlighting mechanisms like skill-biased demand shifts and managerial power expansion that align with the project's themes of labor market winners and losers. The impact of the digital technology application (DTA) on labor income share has sparked intense debate among academics. Nevertheless, few studies have focused on its influence and potential mechanisms on average employee compensation (AEC) and the executive-employee compensation gap (ECG). We employed listed corporations in China from 2011 to 2020 to answer these questions. Our findings are as follows. First, DTA can raise AEC, but it also widens ECG. Second, the potential mechanisms of DTA to enhance AEC are improving gross operating income, optimizing human resource allocation, and boosting total factor productivity. Its potential mechanism for widening ECG also is expanding managerial...	Sai Yuan, Ran Zhou, Mengna Li, et al.
8	2023	The impact of artificial intelligence on labor markets in developing countries: a new method with an illustration for Lao PDR and urban Viet Nam ↗ This paper directly addresses the project's core theme of AI exposure measurement by proposing a novel methodology to translate US-based AI impact scores to developing economies using semantic similarity. It provides empirical evidence on how AI affects labor markets in different occupational structures, contributing to the understanding of distributional effects and task-based frameworks beyond advanced economies. AI is transforming labor markets around the world. Existing research has focused on advanced economies but has neglected developing economies. Different impacts of AI on labor markets in different countries arise not only from heterogeneous occupational structures, but also from the fact that occupations vary across countries in their composition of tasks. We propose a new methodology to translate existing measures of AI impacts that were developed for the US to countries at various levels of economic development. Our method assesses semantic similarities between textual descriptions of work activities in the US and workers' skills elicited in surveys for other countries. We implement the...	Francesco Carbonero, Jeremy Davies, Ekkehard Ernst, et al.
8	2020	Learning Occupational Task-Shares Dynamics for the Future of Work ↗ This paper directly addresses the project's core theme of measuring AI exposure and task-based dynamics by analyzing shifts in occupational task shares using online job postings. It provides relevant empirical evidence on how AI adoption has differentially impacted high, mid, and low-wage occupations, contributing to the understanding of workforce reorganization and future skill requirements. The recent wave of AI and automation has been argued to differ from previous General Purpose Technologies (GPTs), in that it may lead to rapid change in occupations' underlying task requirements and persistent technological unemployment. In this paper, we apply a novel methodology of dynamic task shares to a large dataset of online job postings to explore how exactly occupational task demands have changed over the past decade of AI innovation, especially across high, mid and low wage occupations. Notably, big data and AI have risen significantly among high wage occupations since 2012 and 2016, respectively. We built an ARIMA model to predict future occupational task demands and showcase...	Subhro Das, Sebastian Steffen, Wyatt Clarke, et al.
8	2024	Learning From Ricardo and Thompson: Machinery and Labor in the Early Industrial Revolution and in the Age of Artificial Intelligence ↗ This paper closely relates to the project's themes on AI's distributional effects, wage impacts, and the distinction between productivity growth and worker welfare. It provides a historical analog (Industrial Revolution) to contextualize mechanisms of labor substitution, job degradation, and the conditions under which wages rise, aligning with the project's focus on winners, losers, and wage inequality. David Ricardo initially believed machinery would help workers but revised his opinion, likely based on the impact of automation in the textile industry. Despite cotton textiles becoming one of the largest sectors in the British economy, real wages for cotton weavers did not rise for decades. As E.P. Thompson emphasized, automation forced workers into unhealthy factories with close surveillance and little autonomy. Automation can increase wages, but only when accompanied by new tasks that raise the marginal productivity of labor and/or when there is sufficient additional hiring in complementary sectors. Wages are unlikely to rise when workers cannot push for their share of productivity...	Daron Acemoğlu, Simon Johnson
8	2024	Artificial intelligence's creation and displacement of labor demand ↗ This paper directly addresses the project's core questions regarding the causal effects of AI on labor demand, specifically examining the displacement and creation of jobs across different worker skill levels. It provides empirical evidence on how AI adoption affects both AI-skilled and non-AI workers, offering key insights into the distributional effects and task-based dynamics central to the research project. The paper explores the dynamics of labor demand creation and displacement from adopting artificial intelligence (AI) in US metropolitan statistical areas (MSAs). We combine unique online job postings and patent data to identify AI innovation and AI-skilled labor demand for specific industry sectors and locations. Our analysis shows that AI technologies are increasingly penetrating major industries and disproportionally generating new labor demand for AI-skilled workers in the MSAs in which AI innovation occurs. Our empirical model provides nascent evidence that demand for non-AI labor declines slightly in sectors and MSAs with higher AI skill adoption rates. This decline in labor demand is...	Taelim Choi, Nancey Green Leigh
8	2024	New technologies and jobs in Europe ↗ This paper directly addresses the project's core question regarding the aggregate effects of AI on employment shares across European countries, providing crucial empirical evidence on who the winners and losers are. It further complements the project by analyzing how institutional factors like labor market regulations influence the causal relationship between AI exposure and labor market outcomes. Summary We examine the link between labour market developments and new technologies such as artificial intelligence (AI) and software in 16 European countries over the period 2011–9. Using data for occupations at the three-digit level, we find that on average employment shares have increased in occupations more exposed to AI. This is particularly the case for occupations with a relatively higher proportion of younger and skilled workers. While there exists heterogeneity across countries, only very few countries show a decline in employment shares of occupations more exposed to AI-enabled automation. Country heterogeneity for this result seems to be linked to the pace of technology diffusion...	Stefania Albanesi, António Dias da Silva, Juan F. Jimeno, et al.
8	2023	Is Your Machine Better Than You? You May Never Know ↗ This paper directly addresses the project's theme of human-AI complementarity by modeling how decision-makers supervise and trust AI systems in high-stakes environments. It provides key insights into the behavioral mechanisms of AI adoption and the conditions under which human oversight may fail, which is crucial for understanding how firms reorganize work with AI tools. Artificial intelligence systems are increasingly demonstrating their capacity to make better predictions than human experts. Yet recent studies suggest that professionals sometimes doubt the quality of these systems and overrule machine-based prescriptions. This paper explores the extent to which a decision maker (DM) supervising a machine to make high-stakes decisions can properly assess whether the machine produces better recommendations. To that end, we study a setup in which a machine performs repeated decision tasks (e.g., whether to perform a biopsy) under the DM’s supervision. Because stakes are high, the DM primarily focuses on making the best choice for the task at hand...	Francis de Véricourt, Hüseyin Gürkan
8	2022	Technological and Organizational Change and the Careers of Workers ↗ This paper directly addresses the project's theme of firm reorganization and its distributional effects on workers, particularly regarding the displacement of routine tasks and the specific vulnerability of older workers. It provides crucial evidence on how organizational responses, such as retraining, mediate the impact of technological change on career trajectories and labor market outcomes. This paper investigates the effects of technological and organizational change (T&O) on jobs and workers. We show that although T&O reduces firm demand for routine relative to abstract task-based jobs, affected workers do not face higher probability of non- employment or lower earnings growth than unaffected workers. Rather, firms that adopt T&O offer routine workers re-training opportunities to upgrade to more abstract jobs. Older workers form an important exception: T&O increases the risk that they permanently withdraw from the labor market and reduces their earnings, regardless of the tasks they performed in the firm prior to T&O.	Michèle Battisti, Christian Dustmann, Uta Schönberg
8	2024	Beyond AI Exposure: Which Tasks are Cost-Effective to Automate with Computer Vision? ↗ This paper is closely related to the project's theme of measuring AI exposure, as it critiques existing exposure metrics and proposes a framework integrating technical feasibility and economic cost-effectiveness for automation. It directly addresses how AI impacts labor markets by estimating which tasks are actually likely to be automated, providing crucial context for understanding the pace and distribution of worker displacement. . The faster AI automation spreads through the economy, the more profound its potential impacts, both positive (improved productivity) and negative (worker displacement). The previous literature on “AI Exposure” cannot predict this pace of automation since it attempts to measure an overall potential for AI to affect an area, not the technical feasibility and economic attractiveness of building such systems. In this article, we present a new type of AI task automation model that is end-to-end, estimating: the level of technical performance needed to do a task, the characteristics of an AI system capable of that performance, and the economic choice of whether to build and deploy such a...	Maja Svanberg, Wensu Li, Martin D. Fleming, et al.
8	2023	Are artificial intelligence dividends evenly distributed between profits and wages? Evidence from the private enterprise survey data in China ↗ This paper directly addresses the core question of how AI affects wages and inequality by analyzing the distribution of income between profits and labor demand. It provides relevant empirical evidence on whether AI acts as a substitute for labor, which is central to the project's investigation of winners, losers, and distributional effects. Using artificial intelligence (AI) patents and private enterprise survey data in China, we investigate the impact of AI on the distribution of income between profits and wages. Our results reveal that the impact of AI on the distribution of income between profits and wages is U-shaped, implying that with AI shocks the distribution of income shifts from labor to capital. We find that AI impacts the distribution of income between profits and wages by reducing labor demand and increasing capital productivity. The impact of AI on the distribution of income between profits and wages is heterogeneous across enterprise organizational forms, employee training and dividend incentives, attributes of...	Linhui Wang, Zhanglu Cao, Zhiqing Dong
8	2025	Digital transformation and enterprise employment ↗ This paper directly addresses the project's core theme of AI's causal effects on employment, providing empirical evidence that digital transformation, including AI, shrinks total enterprise employment. It further aligns with distributional analysis by highlighting how these effects differentially impact industries and worker skill levels, specifically increasing the share of highly-skilled labor. Digital transformation refers to the process through which businesses adopt digital technologies to drive change. However, how digital transformation changes the size or the structure of enterprise employment remains unknown. Based on a textual analysis of the annual reports of Chinese A-share listed companies, we develop a new approach to measure the adoption of digital technologies at the enterprise level, and deeply explore the relation between digital transformation and enterprise employment. Overall, our results show that the adoption of digital technologies such as artificial intelligence, big data, cloud computing or blockchain significantly shrinks firms’ total employment, which...	Jie Gao, Zhizhuo Li, Thithuha Nguyen, et al.
8	2023	Labour‐saving automation: A direct measure of occupational exposure ↗ This paper directly addresses the core theme of measuring AI and automation exposure across occupations using a task-based framework. Its development of a fine-grained measure linking robotic patents to ONET task descriptions provides a methodological precedent and relevant context for analyzing how different workers are affected by technological substitution. Abstract This article represents one of the first attempts at building a direct measure of occupational exposure to robotic labour‐saving technologies. After identifying robotic and labour‐saving robotic patents, the underlying 4‐digit CPC (Cooperative Patent Classification) code definitions, together with ONET (Occupational Information Network) task descriptions, are employed to detect functions and operations which are more directed to substituting the labour input and their exposure to labour‐saving automation. This measure allows us to obtain fine‐grained information on tasks and occupations according to their text similarity ranking. Occupational exposure by wage and employment...	Fabio Montobbio, Jacopo Staccioli, Maria Enrica Virgillito, et al.
8	2024	The Uneven Impact of Generative AI on Entrepreneurial Performance ↗ [Title only] The paper directly addresses the core theme of distributional effects by examining how generative AI impacts entrepreneurial performance unevenly. It likely contributes to understanding which workers or sectors are winners and losers, fitting the project's focus on inequality and labor market reshaping. No abstract available.	Nicholas G. Otis, Rowan Philip Clarke, Solène Delecourt, et al.
8	2023	Measuring the Impact of Artificial Intelligence and Robotics on the Workplace ↗ This paper directly addresses the core theme of AI exposure measurement by reviewing and comparing various approaches to assessing automation potential in the workplace. It provides essential methodological context for understanding how AI impacts labor markets, which is central to the researcher's project. Abstract Understanding how AI and robotics impact the workplace is fundamental for understanding the broader impact of these technologies on the economy and society. It can also help in developing realistic scenarios about how jobs and skill demand will be redefined in the next decades and how education systems should evolve in response. This chapter provides a literature review of studies that aim at measuring the extent to which AI and robotics can automate work. The chapter presents five assessment approaches: 1) an approach that focuses on occupational tasks and analyzes whether these tasks can be automated; 2) an approach that draws on information from patents to assess computer...	Mila Staneva, S. D. Elliott
8	2022	Improving Human-Algorithm Collaboration: Causes and Mitigation of Over- and Under-Adherence ↗ This paper directly addresses the core theme of whether AI tools augment or substitute for labor by examining the causal mechanisms of human-algorithm collaboration. It provides empirical evidence on task reorganization and worker productivity, specifically highlighting how cognitive biases affect the effective use of AI recommendations in decision-making tasks. Even if algorithms make better predictions than humans on average, humans may sometimes have “private” information which an algorithm does not have access to that can improve performance. How can we help humans effectively use and adjust recommendations made by algorithms in such situations? When deciding whether and how to override an algorithm’s recommendations, we hypothesize that people are biased towards following a naı̈ve advice weighting (NAW) heuristic: they take a weighted average between their own prediction and the algorithm’s, with a constant weight across prediction instances, regardless of whether they have valuable private information. This leads to humans over-adhering to...	Maya Balakrishnan, Kris Ferreira, Jordan Tong
8	2025	The Cybernetic Teammate: A Field Experiment on Generative AI Reshaping Teamwork and Expertise ↗ [Title only] The title directly addresses the core theme of how AI reshapes work organization and teamwork, aligning with the project's focus on firm-level reorganization and task-based effects. The mention of a field experiment on 'generative AI' and 'expertise' suggests strong empirical relevance for assessing productivity and skill complementarity. No abstract available.	Fabrizio Dell’Acqua, Charles Ayoubi, Hila Lifshitz‐Assaf, et al.
8	2023	Firm Investments in Artificial Intelligence Technologies and Changes in Workforce Composition ↗ This paper directly addresses the project's core themes of firm reorganization, changes in workforce composition, and the impact of AI adoption on hierarchy and skill requirements. It provides empirical evidence on how AI investment shifts hiring towards more educated workers and flattens management structures, offering key insights into the distributional effects and task reorganization aspects of the research agenda. We study the shifts in U.S. firms' workforce composition and organization associated with the use of AI technologies.To do so, we leverage a unique combination of worker resume and job postings datasets to measure firm-level AI investments and workforce composition variables, such as educational attainment, specialization, and hierarchy.We document that firms with higher initial shares of highly-educated workers and STEM workers invest more in AI.As firms invest in AI, they tend to transition to more educated workforces, with higher shares of workers with undergraduate and graduate degrees, and more specialization in STEM fields and IT skills.Furthermore, AI investments are associated with...	Tania Babina, Anastassia Fedyk, Alex Xi He, et al.
8	2025	ChatGPT decreases idea diversity in brainstorming ↗ [Title only] This paper directly addresses the task-based framework and AI skill complementarity by examining how generative AI alters cognitive inputs, specifically idea diversity, which is crucial for creative and high-skill tasks. It provides insights into potential negative externalities of AI on workforce innovation and the quality of output, which relates to the 'winners and losers' question regarding skill-dependent productivity effects. No abstract available.	Lennart Meincke, Gideon Nave, Christian Terwiesch
8	2023	No Great Equalizer: Experimental Evidence on AI in the UK Labor Market ↗ [Title only] The title explicitly references experimental evidence on AI in the labor market, directly addressing the project's core questions regarding causal effects and distributional impacts. Although it does not specify the role of generative AI or task-based frameworks, it aligns strongly with the investigation of winners, losers, and productivity effects in a major economy. No abstract available.	Matthias Haslberger, Jane Gingrich, Jasmine Bhatia
8	2022	The Remainder Effect: How Automation Complements Labor Quality ↗ This paper directly addresses the project's core theme of AI and automation complementing labor by introducing a 'Remainder Effect' where automation raises demand for detailed, unobserved skills. It utilizes a task-based framework to explain how automation affects wages and inequality across diverse occupational groups, aligning closely with the researcher's interest in skill complementarity and distributional effects. Using help-wanted ad data, this paper argues that automation increases demand for detailed skills that are typically unobserved, but which are major determinants of pay. Following automation events, we find that employers request more detailed skills and they substantially increase pay offers (8.7%). Importantly, these increases are not limited to select occupational groups—they apply to both routine and non-routine jobs, to jobs requiring college and those that do not. To explain this phenomenon, we extend the Acemoglu-Restrepo task-based model of automation to consider labor quality, which depends on workers having task-specific skills. We obtain a Remainder Effect: when automation...	James Bessen, Erich Denk, Meng Chen
8	2022	Effects of technological change and automation on industry structure and (wage-)inequality: insights from a dynamic task-based model ↗ This paper directly addresses the project's core themes by using a dynamic task-based model to analyze how automation and AI reshape industry structure, wage inequality, and task allocation. It provides relevant theoretical insights into the distributional effects of AI on workers with different skill levels and the resulting changes in market dynamics. Abstract The advent of artificial intelligence is changing the task allocation of workers and machines in firms’ production processes with potentially wide ranging effects on workers and firms. We develop an agent-based simulation framework to investigate the consequences of different types of automation for industry output, the wage distribution, the labor share, and industry dynamics. It is shown how the competitiveness of markets, in particular barriers to entry, changes the effects that automation has on various outcome variables, and to which extent heterogeneous workers with distinct general skill endowments and heterogeneous firms featuring distinct wage offer rules affect the...	Herbert Dawid, Michael Neugart
8	2024	Diverging paths: AI exposure and employment across European regions ↗ This paper directly addresses the project's theme of distributional effects by analyzing how AI exposure and employment outcomes vary across European regions, highlighting skill and sectoral disparities. It complements the core research question on who the winners and losers are by providing a geographic and structural framework for understanding AI's labor market impacts. • This study explores AI exposure and employment patterns in European regions. • Regional clusters highlight disparities in AI readiness and economic structures. • Innovation, skills and specialisation may shape regional impacts of AI. • In high-tech service and capital centres, AI could potentially complement labour. • AI might deepen regional inequalities, with peripheral areas loosing further ground. This study explores exposure to artificial intelligence (AI) technologies and employment patterns in Europe. First, we provide a thorough mapping of European regions focusing on the structural factors—such as sectoral specialisation, R&D capacity, productivity and workforce skills—that may...	Dario Guarascio, Jelena Reljic, Roman Stöllinger
8	2020	Unlocking the potential of AI for English law ↗ This paper directly addresses the project's themes of AI task substitution, augmentation, and firm reorganization by analyzing how AI reshapes legal work structures and multi-disciplinary teams. It provides specific empirical evidence on how AI adoption affects workforce composition and career ladders within a professional service sector, aligning closely with the research focus on distributional effects and organizational changes. This paper discusses how digital technologies including artificial intelligence (AI) reshape the work of lawyers and the organisations that they work for. We overview how AI is being used in legal services, and identify three distinct impacts: AI substitutes automatable legal tasks; AI enhances productivity of lawyers giving advice on the basis of AI-generated outputs; and legal expertise itself augments the deployment of AI when lawyers work as part of a multi-disciplinary team (MDT) encompassing a range of relevant professional expertise. Our survey of English solicitors shows that AI deployment is associated with MDTs, and that MDTs are less prevalent in law firms than in corporations...	John Armour, Richard Parnham, Mari Sako
8	2025	How does artificial intelligence shock affect labor income distribution? Evidence from China ↗ This paper directly addresses the project's interest in the distributional effects of AI on wages and inequality by analyzing how AI shocks impact labor income shares in China. It provides relevant empirical evidence on skill demand and premium effects, which are key mechanisms for understanding who wins and loses in the AI-driven labor market. Based on the neoclassical growth model and labor-management negotiation framework, we theoretically investigate the impact and mechanism of artificial intelligence (AI) shocks on the labor income share of firms. We then take China's “New-generation Artificial Intelligence Pilot Zone Policy” (AI pilot zone policy) as an exogenous shock and analyze micro-level corporate data. Employing a staggered difference-in-differences model, we find that the AI pilot zone policy significantly increases the labor income share in firms, primarily through the skill demand and skill premium effects. Our results withstand various robustness tests. Furthermore, we observe that the AI pilot zone policy has a...	Xiamin Fan, Yuhui Wu, Yucheng Zhou, et al.
8	2025	The Impact of Artificial Intelligence on Labor Market Income Inequality<b></b> ↗ This paper directly addresses the project's focus on the distributional effects of AI on wages and inequality by outlining specific theoretical channels like skill-biased technical change and capital-labor substitution. It provides relevant background on the mechanisms through which AI may widen the wage gap, aligning closely with the core themes of who the winners and losers are in the AI labor market. As the development and application of Artificial Intelligence (AI) technologies accelerate, their disruptive impact on the labor market has become a central issue in economic research. This paper aims to investigate the multifaceted mechanisms through which AI influences income inequality. It begins by reviewing the established theoretical frameworks linking technological progress to income distribution, followed by a survey of contemporary literature on AI's labor market effects. Building on this foundation, the paper develops a theoretical framework that delineates four primary channels through which AI may exacerbate inequality: enhanced skill-biased technical change (SBTC)...	Yongxiang Gu, Yanhui Wang
8	2023	Does artificial intelligence kill employment growth: the missing link of corporate AI posture ↗ The paper directly addresses the project's core question regarding the causal effects of AI on employment by analyzing firm-level AI adoption and its impact on employment growth. It contributes relevant empirical context by distinguishing between AI automation and innovation postures, offering insights into how firms reorganize work and the distributional outcomes for workers. Introduction: An intense debate has been on-going about how artificial intelligence (AI) technology investments have an impact on employment. The debate has often focused on the potential of AI for human task automation, omitting the strategic incentive for firms to cooperate with their workers as to exploit AI technologies for the most relevant benefit of new product and service innovation. Method: We calibrate an empirical probit regression model of how changes in employment relate to AI diffusion, based on formalizing a game-theoretical model of a firm exploiting the twin role of AI innovation and AI automation for both absolute and competitive advantage. Results: The theoretical...	Jacques Bughin
8	2023	Who is AI Replacing? The Impact of ChatGPT on Online Freelancing Platforms ↗ [Title only] This paper directly addresses the core theme of distributional effects and substitution by investigating how AI impacts employment within online labor markets, a key dataset for this project. It specifically tackles the question of whether AI tools replace or augment labor for specific worker groups, providing causal evidence on the 'winners and losers' of AI adoption. No abstract available.	Ozge Demirci, Jonas Hannane, Xinrong Zhu
8	2023	GPTs and Labor Markets in the Developing Economy: Evidence from China ↗ This paper directly addresses the core theme of measuring AI exposure across occupations and regions using a task-based framework. It provides valuable empirical context on how GPT capabilities map to labor markets in a major developing economy, complementing existing U.S.-focused studies. In this paper, we investigate the potential impact of recent advances in AI, Generative Pre-trained Transformers (GPTs), on the labor markets in China. Using occupational data and the census data about the Chinese economy, we first adopt a recently developed methodology and design a rubric to link China’s occupations with new capabilities enabled by GPTs. Then we systematically estimate the extent to which occupations, industries, and geographic regions are exposed to the capabilities of GPTs in China. Our findings show that the contemporary labor markets in China have lower potential of exposure to GPTs compared to the U.S., in terms of the percentage of tasks within occupations exposed to...	Bowen Lou, Hongshen Sun, Tianshu Sun
8	2024	When Advanced AI Isn't Enough: Human Factors as Drivers of Success in Generative AI-Human Collaborations ↗ This paper directly addresses the project's core themes of AI skill complementarity, distributional effects, and the causal impact of AI on worker productivity through rigorous experimental methods. It provides critical empirical evidence that AI adoption does not act as an equalizer but rather exacerbates disparities based on AI literacy, thereby informing the study of winners and losers in the AI-driven labor market. In this comprehensive study, we explore the dynamics of human-AI collaboration through two randomized controlled experiments, focusing on the role of generative AI and its interaction with humans. Our investigation demonstrates that access to generative AI significantly enhances performance outcomes, highlighting its importance as a performance determinant. However, our findings challenge the notion of AI as a great equalizer; while AI usage leads to improved performance, it does not necessarily compress variance among individuals, indicating the emergence of new skill disparities in the AI era. We found that working with advanced AI models, such as GPT-4.0, only slightly improves...	Ning Li, Huaikang Zhou, Wenming Deng, et al.
8	2023	The Unintended Consequences of Censoring Digital Technology - Evidence from Italy's ChatGPT Ban ↗ This paper directly investigates the causal effects of AI on worker productivity by analyzing the output of software developers before and after a specific AI tool ban. It provides valuable empirical evidence on how AI adoption impacts individual performance and highlights the immediate disruptions caused by restricting access to generative AI tools. We analyse the effects of the ban of ChatGPT, a generative pre-trained transformer chatbot, on individual productivity. We first compile data on the hourly coding output of over 8,000 professional GitHub users in Italy and other European countries to analyse the impact of the ban on individual productivity. Combining the high-frequency data with the sudden announcement of the ban in a difference-in-differences framework, we find that the output of Italian developers decreased by around 50\% in the first two business days after the ban and recovered after that. Applying a synthetic control approach to daily Google search and Tor usage data shows that the ban led to a significant increase in...	David Kreitmeir, Paul A. Raschky
8	2022	Ai, Skill, and Productivity: The Case of Taxi Drivers ↗ This paper directly addresses the core question of whether AI tools augment or substitute for labor by demonstrating that AI assistance acts as a substitute for low-skilled taxi drivers' route-finding abilities. It provides empirical evidence on the distributional effects of AI by showing how productivity gains are accrued only to low-skilled workers, thereby narrowing the productivity gap between high- and low-skilled drivers. We examine the impact of artificial intelligence (AI) on productivity in the context of taxi drivers. The AI we study assists drivers with finding customers by suggesting routes along which the demand is predicted to be high. We find that AI improves drivers’ productivity by shortening the cruising time, and this gain is accrued only to low-skilled drivers, narrowing the productivity gap between high- and low-skilled drivers by 13.4%. This case study provides evidence that AI and skill are indeed substitutes, offering direct support for the underlying assumption of recent projection exercises regarding job displacement by AI. This paper was accepted by Joshua Gans, business strategy...	Kyogo Kanazawa, Daiji Kawaguchi, Hitoshi Shigeoka, et al.
8	2023	<div> AI-enabled Technology and Gig Workforce: The Role <span>of Experience, Skill Level, and Task Complexity</span></div> ↗ [Title only] This paper directly addresses the core themes of online labor markets, task complexity, and distributional effects by examining how AI interacts with gig workers' skills and experience. It provides specific insights into whether AI tools augment or substitute for labor based on task characteristics, which is central to the project's inquiry into labor market reshaping. No abstract available.	Benjamin Knight, Dmitry Mitrofanov, Serguei Netessine
8	2024	The Impact of Generative Artificial Intelligence on Socioeconomic Inequalities and Policy Making ↗ [Title only] The title directly addresses the distributional effects and inequality aspects of AI, which are core themes of the research project. Although the focus on 'policy making' suggests a broader scope than pure labor economics, the explicit mention of socioeconomic inequalities aligns closely with the investigation of winners and losers across different worker demographics. No abstract available.	Valerio Capraro, Austin Lentsch, Daron Acemoğlu, et al.
8	2024	Can LLMs Answer Investment Banking Questions? Using Domain-Tuned Functions to Improve LLM Performance on Knowledge-Intensive Analytical Tasks ↗ This paper directly investigates the productivity and accuracy of LLMs in a high-skill, knowledge-intensive occupation, addressing the project's core themes of task-based AI exposure and labor augmentation. It provides empirical evidence on whether AI tools substitute for or complement specialized analytical skills, which is central to understanding the distributional effects of AI on professional workers. Large Language Models (LLMs) can increase the productivity of general-purpose knowledge work, but accuracy is a concern, especially in professional settings requiring domain-specific knowledge and reasoning. To evaluate the suitability of LLMs for such work, we developed a benchmark of 16 analytical tasks representative of the investment banking industry. We evaluated LLM performance without special prompting, with relevant information provided in the prompt, and as part of a system giving the LLM access to domain-tuned functions for information retrieval and planning. Without access to functions, state-of-the-art LLMs performed poorly, completing two or fewer tasks correctly. Access to...	Nicholas Harvel, Felipe Bivort Haiek, Anupriya Ankolekar, et al.
8	2025	Generation AI: Job Crafting by Entry-Level Professionals in the Age of Generative AI ↗ This paper directly addresses the core theme of how entry-level workers respond to AI adoption by examining job crafting behaviors and the emergence of 'signal crafting' in knowledge work. It provides nuanced insights into task reorganization and career prospects for entry-level professionals, which are central to understanding the distributional effects and labor market impacts of generative AI. Abstract The rise of Generative Artificial Intelligence (GenAI) in the workplace is transforming knowledge work in organizations with important implications for professionals and organizations alike. This study focuses on entry-level professionals in knowledge work (ELPs) and highlights how ELPs reshape their work in response to the adoption and integration of GenAI in the workplace. Drawing on the theoretical lens of job crafting and building on insights from a qualitative study in consultancy, this research shows how ELPs use GenAI to proactively (re)craft their work tasks and relationships. Moreover, the study identifies a third behavioral dimension of job crafting that is becoming...	Anne‐Sophie Mayer, Reza Mousavi, Reinout Buwalda
8	2023	Six questions about the demand for artificial intelligence skills in labour markets ↗ This paper directly addresses the project's core theme of measuring AI exposure across occupations by analyzing the diffusion of AI skill demands in job postings. It provides crucial empirical context on how AI skills interact with wage premiums and other cognitive skills, informing the study of labor market inequality and skill complementarity. This study responds to six key questions about the impact that the demand for Artificial Intelligence (AI) skills is having on labour markets. What are the occupations where AI skills are most relevant? How do different AI-relevant skills combine in job requirements? How quickly is the demand for AI-related skills diffusing across labour markets and what is the relationship between AI skill demands and the demand for cognitive skills across jobs? Finally, are AI skills leading to a wage premium and how different are the wage returns associated with AI and routine skills? To shed light on these aspects, this study leverages Natural Language Processing (NLP) algorithms to analyse the...	Fabio Manca
8	2022	Automation and occupational mobility: A task and knowledge-based approach ↗ This paper directly addresses the core themes of automation's impact on labor and occupational mobility by proposing a task-based framework to analyze how workers adapt to displacement. It complements the project's inquiry into distributional effects and labor market adjustments by highlighting the role of task similarity in facilitating occupational transitions. How does automation affect labour? Academic literature emphasises that automation leads to job displacement, polarisation of labour, slower wage growth for the middle skilled and more. However, existing literature seldom discusses ways individuals could adapt to automation. One such insufficiently explored adaptation strategy is occupational mobility. To fill this gap, this article proposes and validates a task and knowledge based occupational mobility network that takes into account automation. The result of the analysis shows that many compelling insights can be derived from such a network. First, many occupations cluster together with similar automation probabilities, though some...	Aleksandr Christenko
8	2024	Experimenting with Generative AI: Does ChatGPT Really Increase Everyone’s Productivity? ↗ [Title only] This title directly addresses the core theme of generative AI productivity experiments and investigates whether productivity gains are universal or heterogeneous across workers. It aligns closely with the project's interest in causal effects on worker productivity and the distributional winners and losers of AI adoption. No abstract available.	Voraprapa Nakavachara, Tanapong Potipiti, Thanee Chaiwat
8	2025	Expertise ↗ This paper directly addresses the project's core question of whether technology augments or substitutes labor by introducing an 'expertise' framework that explains heterogeneous wage and employment effects. It provides a relevant task-based mechanism and empirical evidence on how automation reshapes labor market dynamics, aligning closely with the themes of skill complementarity, distributional effects, and task reorganization. Abstract When job tasks are automated, does this augment or diminish the value of labor in the tasks that remain? We argue the answer depends on whether removing tasks raises or reduces the expertise required for remaining non-automated tasks. Since the same task may be relatively expert in one occupation and inexpert in another, automation can simultaneously replace experts in some occupations while augmenting expertise in others. We propose a conceptual model of occupational task bundling that predicts that changing occupational expertise requirements have countervailing wage and employment effects: automation that decreases expertise requirements reduces wages but permits the entry of...	David Autor, Neil Thompson
8	2025	Artificial Intelligence in the Knowledge Economy ↗ This paper provides a theoretical framework linking AI autonomy levels to labor market outcomes within hierarchical organizations, directly addressing the project's inquiry into whether AI augments or substitutes labor based on skill level. It offers key mechanisms for understanding distributional effects and firm reorganization, particularly regarding the impact on entry-level workers versus expert solvers. Artificial Intelligence (AI) can transform the knowledge economy by automating non-codifiable work. To analyze this transformation, we incorporate AI into an economy where humans form hierarchical organizations: Less knowledgeable individuals become"workers"doing routine work, while others become"solvers"handling exceptions. We model AI as a technology that converts computational resources into"AI agents"that operate autonomously (as co-workers and solvers/co-pilots) or non-autonomously (solely as co-pilots). Autonomous AI primarily benefits the most knowledgeable individuals; non-autonomous AI benefits the least knowledgeable. However, output is higher with autonomous AI. These findings...	Enrique Ide, Eduard Talamàs
8	2025	Shifting Work Patterns with Generative AI ↗ This paper provides direct causal evidence on how generative AI adoption alters task performance and time allocation for knowledge workers, addressing the project's core interest in AI's impact on productivity and work reorganization. The randomized field experiment design offers high-quality empirical insight into the mechanisms of AI-augmented labor, specifically regarding individual-level behavioral changes in email and document creation tasks. We present evidence on how generative AI changes the work patterns of knowledge workers using data from a 6-month-long, cross-industry, randomized field experiment.Half of the 7,137 workers in the study received access to a generative AI tool integrated into the applications they already used for emails, document creation, and meetings.We find that access to the AI tool during the first year of its release primarily impacted behaviors that workers could change independently and not behaviors that require coordination to change: workers who used the tool in more than half of the sample weeks spent 3.6 fewer hours, or 31% less time on email each week (intent to treat estimate is 1.3 hours)...	Eleanor Wiske Dillon, Sonia Jaffe, Nicole Immorlica, et al.
8	2025	AI, Skill, and Productivity: The Case of Taxi Drivers ↗ This paper directly addresses the project's core theme of whether AI tools augment or substitute for labor by providing empirical evidence that predictive AI acts as a substitute for low-skilled taxi drivers. It offers valuable insights into distributional effects, specifically showing how AI adoption narrows productivity gaps between high- and low-skilled workers, which aligns with the research focus on inequality and task-based analysis. We examine the impact of artificial intelligence (AI) on productivity in the context of taxi drivers. The AI we study assists drivers with finding customers by suggesting routes along which the demand is predicted to be high. We find that AI improves drivers’ productivity by shortening the cruising time, and this gain is accrued only to low-skilled drivers, narrowing the productivity gap between high- and low-skilled drivers by 13.4%. This case study provides evidence that AI and skill are indeed substitutes, offering direct support for the underlying assumption of recent projection exercises regarding job displacement by AI. This paper was accepted by Joshua Gans, business strategy...	Kyogo Kanazawa, Daiji Kawaguchi, Hitoshi Shigeoka, et al.
8	2025	Artificial Intelligence, Technological Innovation, and Employment Transformation for Sustainable Development: Evidence from China ↗ This paper directly addresses the distributional effects of AI on labor markets by analyzing how AI adoption differentially impacts employment demands across skill levels and genders in China. It provides relevant empirical evidence on task-based substitution and complementarity, specifically highlighting the suppression of low-skilled labor and augmentation of high-skilled labor, which aligns with the project's core themes of inequality and skill complementarity. With the rapid advancement of artificial intelligence (AI) technology, the global employment structure is undergoing profound transformations, significantly impacting social sustainability. This study utilizes panel data from 30 Chinese provinces spanning the years 2010 to 2022 and applies a two-way fixed-effects model to analyze the impact of AI development on the employment skills structure. The findings indicate that advancements in AI technology significantly suppress the demand for low-skilled labor while markedly enhancing the demand for both middle- and high-skilled labor. The threshold effect analysis reveals a nonlinear relationship between AI advancements and the demand for...	Hui Liang, Jingbo Fan, Yunhan Wang
8	2024	Generative AI, Work and Risks in Cultural and Creative Industries ↗ [Title only] This title directly addresses the intersection of generative AI and labor markets within a specific sector, aligning with the project's interest in distributional effects and firm reorganization. However, the focus on 'risks' and 'cultural industries' may limit its direct applicability to broader aggregate labor market questions or general productivity experiments. No abstract available.	Emmanuelle Walkowiak, Jason Potts
8	2024	How AI Outperforms Humans at Creative Idea Generation ↗ [Title only] This paper directly addresses the core theme of AI augmenting or substituting labor within the specific domain of creative tasks, which is central to understanding task-based framework shifts. By providing empirical evidence on AI's superior performance in idea generation, it offers critical insights into how generative AI reshapes occupational productivity and the distribution of winners and losers. No abstract available.	Noah Castelo, Zsolt Katona, Peiyao Li, et al.
8	2025	Artificial Intelligence, Scientific Discovery, and Product Innovation ↗ This paper directly addresses the project's core theme of AI skill complementarity by demonstrating how AI augments top scientists' productivity while substituting routine idea-generation tasks. It provides crucial empirical evidence on distributional effects and task reorganization, highlighting the disparity in benefits across skill levels and the potential trade-offs between productivity and job satisfaction. This paper studies the impact of artificial intelligence on innovation, exploiting the randomized introduction of a new materials discovery technology to 1,018 scientists in the R amp;D lab of a large U.S. firm. AI-assisted researchers discover 44% more materials, resulting in a 39% increase in patent filings and a 17% rise in downstream product innovation. These compounds possess more novel chemical structures and lead to more radical inventions. However, the technology has strikingly disparate effects across the productivity distribution: while the bottom third of scientists see little benefit, the output of top researchers nearly doubles. Investigating the mechanisms behind these...
8	2024	Investigating employees’ occupational risks and benefits resulting from artificial intelligence: An empirical analysis ↗ This paper directly addresses the project's core question of how AI exposure varies across workers and occupations by proposing a refined framework for measuring AI benefits and risks. It aligns with the themes of distributional effects and task-based analysis by examining how skills, education, and demographics correlate with these outcomes. With rapid advances in artificial intelligence (AI), more employees are benefiting from or being replaced by AI. Nevertheless, we know little about the extent to which AI affects employees’ occupations positively. This study improves the methodologies for quantifying employees’ occupational AI benefits and risks. We propose three mechanisms by which AI may benefit employees’ careers: productivity-enhanced AI jobs, intelligence-augmented AI jobs, and AI-enabling jobs. We also conduct employee-level analyses regarding how employees’ skills, educational backgrounds, and demographics may correlate with occupational risk and AI benefits. Our results suggest that these key factors have distinct...	Qi Wang, Xuanqi Liu, Ke‐Wei Huang
8	2025	AI exposure predicts unemployment risk: A new approach to technology-driven job loss ↗ This paper directly addresses the project's core question regarding the causal effects of AI exposure on employment by developing a novel metric for occupation-level unemployment risk. It provides critical empirical evidence on how AI adoption correlates with job separations, offering key insights into the distributional effects and displacement dynamics central to the research agenda. Is AI disrupting jobs and creating unemployment? This question has stirred public concern for job stability and motivated studies assessing occupations' automation risk. These studies used readily available employment and wage statistics to quantify occupational changes for employed workers. However, they did not directly examine unemployment dynamics primarily due to the lack of data across occupations, geography, and time. Here, we overcome this barrier using monthly occupation-level unemployment data from each US state's unemployment insurance office from 2010 to 2020 to assess AI exposure models, job separations, and unemployment through a new measure called unemployment risk. We...	Morgan R. Frank, Yong‐Yeol Ahn, Esteban Moro
8	2024	The Adoption of ChatGPT ↗ [Title only] The title directly addresses the core theme of firm AI adoption and likely explores how organizations integrate large language models, which is central to the project. It may provide evidence on task reorganization or worker productivity effects, though the specific focus on distributional or aggregate outcomes remains uncertain without more detail. No abstract available.	Anders Humlum, Emilie Vestergaard
8	2025	Generative AI and the augmentation of information practices in knowledge work ↗ [Title only] This title directly addresses the project's core theme of whether AI tools augment or substitute for labor within knowledge work contexts. It likely provides critical evidence on task reorganization and productivity effects for specific occupations, aligning closely with the investigation of generative AI's impact on worker tasks. No abstract available.	Mohammad Hossein Jarrahi, Lan Li, Andrew Phillip Robinson, et al.
8	2023	Artificial Intelligence versus Software Engineers: An Evidence-Based Assessment Focusing on Non-Functional Requirements ↗ This paper directly addresses the core themes of AI substitution versus augmentation and task-based labor market effects by empirically comparing AI and human software engineers on specific coding tasks. It provides relevant evidence on how AI tools reshape work organization and productivity within a high-skill occupation, informing questions about skill complementarity and the future of specialized labor. <title>Abstract</title> The automation of Software Engineering (SE) tasks using Artificial Intelligence (AI) is growing, with AI increasingly leveraged for project management, modeling, testing, and development. Notably, ChatGPT, an AI-powered chatbot, has been introduced as a versatile tool for code writing and test plan generation. Despite the excitement around AI's potential to elevate productivity and even replace human roles in software development, solid empirical evidence remains scarce. Normally, a software engineer's solution is evaluated against a variety of non-functional requirements such as performance, efficiency, reusability, and usability, among others. This study presents...	Nathalia Nascimento, Paulo Alencar, Donald Cowan
8	2023	Generative Ai and Firm Values ↗ This paper directly addresses the project's core themes of measuring AI exposure at the firm level and investigating whether AI substitutes or complements labor. It provides empirical evidence on the causal effects of Generative AI on firm value and profitability, which informs the broader questions regarding distributional effects and labor market reorganization. How do recent advances in Generative AI affect firm value? We construct the first measure of firms’ workforce exposures to Generative AI and show that an “Artificial-Minus-Human” (AMH) portfolio earned 5% in the two weeks following the release of ChatGPT. The labor-exposure effect is more pronounced for firms with greater data assets and is distinct from the effect of firms’ product exposures to AI. We assess whether exposed workforces are substituted or complemented by Generative AI based on whether their exposed tasks are core or supplemental. Examining firms’ labor demand and profitability following the release of ChatGPT supports a labor-technology substitution channel.	Andrea L. Eisfeldt, Gregor Schubert, Miao Ben Zhang
8	2024	Generative AI Enhances Team Performance and Reduces Need for Traditional Teams ↗ This paper directly addresses the core theme of how firms and individuals reorganize work in response to AI adoption by examining shifts in team dynamics and structure. It provides empirical evidence on whether AI acts as an augmentative tool or a substitute for traditional labor collaboration, which is central to understanding the causal effects of AI on worker productivity. Recent advancements in generative artificial intelligence (AI) have transformed collaborative work processes, yet the impact on team performance remains underexplored. Here we examine the role of generative AI in enhancing or replacing traditional team dynamics using a randomized controlled experiment with 435 participants across 122 teams. We show that teams augmented with generative AI significantly outperformed those relying solely on human collaboration across various performance measures. Interestingly, teams with multiple AIs did not exhibit further gains, indicating diminishing returns with increased AI integration. Our analysis suggests that centralized AI usage by a few team...	Ning Li, Huaikang Zhou, Kris Mikel-Hong
8	2025	Large language models at work in China’s labor market ↗ This paper directly addresses the project's core themes by measuring AI exposure in a major labor market and analyzing its distributional effects on wages and experience premiums. It contributes to the project's inquiry into whether AI tools augment or substitute for labor by proposing a novel theoretical framework that deviates from traditional routinization hypotheses. This paper explores the potential impacts of large language models (LLMs) on the Chinese labor market. We analyze occupational exposure to LLM capabilities by incorporating human expertise and LLM classifications, following the methodology of Eloundou et al. (2023). The results indicate a positive correlation between occupational exposure and both wage levels and experience premiums at the occupation level. This suggests that higher-paying and experience-intensive jobs may face greater exposure risks from LLM-powered software. We then aggregate occupational exposure at the industry level to obtain industrial exposure scores. Both occupational and industrial exposure scores align with expert...	Qin Chen, Jinfeng Ge, Huaqing Xie, et al.
8	2024	AI, Task Changes in Jobs, and Worker Reallocation ↗ [Title only] The title directly addresses the core themes of task-based frameworks and worker reallocation, which are central to understanding how AI reshapes labor markets. It likely investigates the causal mechanisms of job restructuring and displacement, offering high relevance to questions about who the winners and losers are. No abstract available.	Christina Gathmann, Felix Grimm, Erwin Winkler
8	2024	Developers’ Perceptions on the Impact of Chatgpt in Software Development: A Survey ↗ This paper is closely related as it investigates the impact of Generative AI on software developers, a key occupation and sector within the project's scope. It addresses core themes such as AI exposure, productivity effects, and worker perceptions of job displacement and task reorganization. As Large Language Models (LLMs), including ChatGPT and analogous systems, continue to advance, their robust natural language processing capabilities and diverse applications have garnered considerable attention. Nonetheless, despite the increasing acknowledgment of the convergence of Artificial Intelligence (AI) and Software Engineering (SE), there is a lack of studies involving the impact of this convergence on the practices and perceptions of software developers. Understanding how software developers perceive and engage with AI tools, such as ChatGPT, is essential for elucidating the impact and potential challenges of incorporating AI-driven tools in the software development process. In...	Thiago Santana Vaillant, Felipe Deveza de Almeida, Paulo Sérgio Panse Silveira, et al.
8	2025	Generative AI Adoption and Higher Order Skills ↗ [Title only] This title directly addresses the intersection of generative AI and skill demand, a core theme regarding AI's impact on labor markets and worker composition. It suggests an investigation into how AI adoption shifts the premium on higher-order skills, which is critical for understanding distributional effects and the augmentation versus substitution debate. No abstract available.	Piyush Gulati, Arianna Marchetti, Phanish Puranam, et al.
8	2024	Generative AI and the Future of Work: Augmentation or Automation? ↗ This paper directly addresses the project's core theme of whether AI augments or substitutes labor, offering a strategic perspective on firm reorganization and the necessity of human oversight. It provides valuable context on sectoral exposure and policy implications, though it is a conceptual report rather than an empirical study of causal labor market effects. 1 This report examines the potential impact of Generative artificial intelligence (AI) systems, such as ChatGPT, on the future of work and, by implication, on productivity. It argues that although Generative AI is powerful, it has significant limitations and risks that require humans to remain “in the loop” not only to prevent systems from going off the rails, but to capture value. Rather than taking a deterministic view that artificial intelligence (AI) will inevitably destroy jobs, the article suggests that an analysis should start with how firms can strategically deploy these tools to gain an advantage. It asks whether “augmentation” or “simplistic automation” lies ahead. Our objective...	John Zysman, Mark Nitzberg
8	2024	Artificial Intelligence in the Knowledge Economy ↗ This paper directly addresses the project's interest in how AI reshapes firm organization and labor markets through a theoretical framework of hierarchical firms and task-based substitution. It provides key insights into the distributional effects of AI by distinguishing between basic and advanced AI, and analyzing their divergent impacts on worker productivity and knowledge content. This paper provides a new framework for studying the impact of Artificial Intelligence (AI) on the organization of knowledge work. We incorporate AI into an economy where humans endogenously form hierarchical firms: Less knowledgeable agents become "workers" solving routine problems, while more knowledgeable agents become "solvers" handling exceptions. We model AI as an algorithm that uses compute to mimic humans. We compare the equilibrium before and after AI's introduction, distinguishing between "basic" AI (with knowledge equivalent to pre-AI workers) and "advanced" AI (with knowledge equivalent to pre-AI solvers). We show that basic AI increases the knowledge content of human work...	Enrique Ide, Eduard Talamàs
8	2024	Corporate Responses to Generative AI: Early Evidence from Conference Calls ↗ [Title only] This paper directly addresses the core theme of firm reorganization and AI adoption by analyzing corporate discourse for early signals of generative AI integration. It provides qualitative evidence on how firms perceive and discuss AI, which is crucial for understanding the initial stages of labor market transformation before hard employment data becomes available. No abstract available.	Ning Jia, Ningzhong Li, Guang Ma, et al.
8	2025	Can a troubleshooting AI assistant improve task performance in industrial contexts? ↗ This paper directly addresses the project's theme of how AI tools augment labor by providing empirical evidence of productivity gains for less experienced workers, highlighting AI's role in bridging skill gaps. It complements the research on distributional effects by demonstrating that AI assistance can enhance performance in complex industrial tasks, particularly for lower-skill or entry-level technicians. Access to domain expertise is critical to problem-solving activities. This study investigates the role and usefulness of an AI-based troubleshooting assistant in the knowledge-intensive context of train commissioning. The authors developed a multilingual, user-centred chatbot using a Retrieval-Augmented Generation framework and a Large Language Model to provide real-time, project-specific troubleshooting support. A controlled field experiment with 19 commissioning technicians completing 173 tasks was conducted to evaluate the effect of the AI assistant. Results show that AI-assisted users significantly outperformed non-users in task performance. The benefits were more substantial among less...	Nils Löwhagen, Philippe Schwendener, Torbjørn H. Netland
8	2025	The Cybernetic Teammate: A Field Experiment on Generative Ai Reshaping Teamwork and Expertise ↗ [Title only] The paper directly addresses the project's core theme of how AI tools reorganize work and reshape teamwork through a field experiment on generative AI. It likely provides causal evidence on whether AI augments or substitutes for labor, specifically examining impacts on expertise and collaborative dynamics. No abstract available.	Fabrizio Dell’Acqua, Charles Ayoubi, Hila Lifshitz, et al.
8	2025	Human after all: Occupations at the core of AI adoption ↗ This paper closely addresses firm-level AI adoption by identifying ICT engineers as the critical occupational driver for implementing AI technologies. It provides relevant empirical evidence on how specific internal skill structures influence the diffusion of AI within organizations, aligning with the project's focus on firm reorganization and skill complementarity. This paper investigates how firms’ occupational structure shapes the adoption of artificial intelligence (AI) using matched administrative data on French firms and relying on an instrumental variable Probit model. We identify ICT engineers as the only occupational group with a robust and statistically significant effect on AI adoption. This finding holds for ICT and non-ICT Services sectors, and regardless of whether AI is developed in-house or acquired externally. Our estimates suggest that closing the occupational gap between adopters and non-adopters would require approximately 215,000 additional ICT engineers, and 45,000 for the firms most exposed to AI. The results highlight the...	Luca Fontanelli, Flavio Calvino, Chiara Criscuolo, et al.
8	2024	De-routinization in the Fourth Industrial Revolution – Firm-Level Evidence ↗ [Title only] The title directly addresses the core theme of task reorganization by focusing on de-routinization, a key mechanism for understanding how AI augments or substitutes labor. However, the absence of 'AI' in the title suggests the paper may focus on broader industrial automation rather than specific machine learning or LLM effects, introducing uncertainty regarding its direct relevance to the project's technological focus. No abstract available.	Melanie Arntz, Sabrina Genz, Terry Gregory, et al.
8	2025	Tech for Good: The employment effects of policy-driven artificial intelligence development ↗ This paper directly addresses the project's core question regarding the causal effects of AI on aggregate employment by exploiting a policy-driven quasi-natural experiment. It provides valuable evidence on distributional outcomes by analyzing heterogeneity across skill levels and firm characteristics, aligning with the project's interest in winners and losers. Employment stability constitutes a fundamental pillar of socioeconomic development, with direct implications for social welfare and societal cohesion. This study utilizes the implementation of Artificial Intelligence Innovation Development Pilot Zones (AIIDPZs) initiated in 2019 as a quasi-natural experiment to investigate how artificial intelligence (AI) development influences corporate employment decisions. Our findings demonstrate that the establishment of AIIDPZs promotes enterprise labor employment, with firms in pilot regions experiencing an average increase in employment scale of approximately 4.56 % compared to enterprises in non-pilot regions, an effect similarly verified at the...	Fei Gao, Benhong Peng, Yinyin Zhao, et al.
8	2025	The impact of AI adoption on R&D productivity: Evidence from Chinese pharmaceutical manufacturing industry ↗ This paper directly examines the causal effect of AI adoption on worker productivity within a specific R&D-intensive sector, addressing the core theme of generative AI productivity experiments. It provides empirical evidence on how AI tools augment labor by enabling firms to reallocate human resources toward higher-skilled 'elite' researchers, offering insights into task reorganization and skill complementarity. The decline in R&D productivity is a persistent and widespread phenomenon hindering long-term economic growth. With the boom of artificial intelligence (AI), especially the pervasive application of deep learning, AI is now able to participate in innovation, which is expected to help reverse the decreasing trend of R&D productivity. Focusing on the pharmaceutical manufacturing industry, one of the most active domains in AI-driven innovation, we take Chinese listed firms as an example to investigate the role of AI in drug discovery and the impact of AI adoption on new drug R&D productivity. Our empirical results show that other things being equal, new drug output per billion yuan invested in...	Yifan Wu, Yiming Yuan, Xueyin Song
8	2023	The Layoff Generation: How Generative Ai Will Reshape Employment and Labor Markets ↗ [Title only] The title explicitly addresses the impact of generative AI on employment and labor markets, directly aligning with the project's core question regarding causal effects on employment and the distributional consequences for workers. Although the focus is broad rather than task-specific, it strongly resonates with themes of AI substitution effects and aggregate labor market outcomes. No abstract available.	Qiwen Huang, Yifan Shen, Yuanchi Sun, et al.
8	2020	Super Mario Meets AI: The Effects of Automation on Team Performance and Coordination in a Videogame Experiment ↗ This paper directly addresses the core theme of how AI reshapes labor markets by examining the causal effects of automation on team performance, coordination, and trust. It provides valuable experimental evidence on AI skill complementarity and substitution, specifically highlighting adverse spillovers and reduced effort in low- and medium-skilled teams. Recent advances in artificial intelligence (AI) have piqued interest in how these technological advances will transform jobs and labor markets. While prior work has focused on understanding the tasks where AI outperforms humans, we ask how the introduction of automated agents affects teams, their routines, and organizations. We randomize the introduction of automated agents and new hires into "experimental firms" engaging in a coordination-based game on the Nintendo Switch console. We demonstrate experimentally that even in a task where automated agents outperform humans, the introduction of an automated agent decreases team performance. These effects are especially large in the short-term...	Fabrizio Dell’Acqua, Bruce Kogut, Patryk Perkowski
8	2024	Mirror, Mirror on the Wall: Which Jobs Will AI Replace After All?: A New Index of Occupational Exposure ↗ This paper directly addresses the core theme of measuring AI exposure by introducing a novel index (GENOE) that quantifies the potential impact of AI on occupations and tasks. It provides relevant methodological context for the project's central question of how to measure AI exposure, particularly by leveraging LLMs to assess job replacement likelihood. This paper introduces the AI Generated Index of Occupational Exposure (GENOE), a novel measure quantifying the potential impact of artificial intelligence on occupations and their associated tasks. Our methodology employs synthetic AI surveys, leveraging large language models to conduct expert-like assessments. This approach allows for a more comprehensive evaluation of job replacement likelihood, minimizing human bias and reducing assumptions about the mechanisms through which AI innovations could replace job tasks and skills. The index not only considers task automation, but also contextual factors such as social and ethical considerations and regulatory constraints that may affect the...	Miguel Benítez, Eric Parrado
8	2025	Artificial Intelligence and the Philippine Labor Market ↗ This paper directly addresses the project's core theme of measuring AI exposure across occupations and workers using a task-based framework of complementarity. It provides valuable empirical evidence on distributional effects by skill, age, and sector, specifically highlighting the BPO industry as a case study for displacement risks. This paper combines labor force survey microdata with measures of occupational AI exposure and complementarity to examine the potential impact of recent advancements in AI on the Philippine labor market. We find that around one third of workers are highly exposed to AI with around sixty percent of those also rated highly complementary, indicating potential productivity gains. College-educated, young, urban, female, and well-paid workers in the services sector are most exposed. Business process outsourcing (BPO) is identified as the sector with the highest proportion of jobs at risk of displacement. Addressing regulatory gaps, infrastructure needs, and workforce reskilling is crucial to...	Micholo Cucio, Tristan Hennig
8	2024	Learning from Ricardo and Thompson: Machinery and Labor in the Early Industrial Revolution, and in the Age of AI ↗ This paper directly addresses the project's core question regarding whether AI augments or substitutes for labor by drawing historical parallels to the Industrial Revolution. It provides critical context on the distributional effects of automation, specifically focusing on wage dynamics, job quality degradation, and the conditions under which productivity gains translate to worker welfare. David Ricardo initially believed machinery would help workers but revised his opinion, likely based on the impact of automation in the textile industry. Despite cotton textiles becoming one of the largest sectors in the British economy, real wages for cotton weavers did not rise for decades. As E.P. Thompson emphasized, automation forced workers into unhealthy factories with close surveillance and little autonomy. Automation can increase wages, but only when accompanied by new tasks that raise the marginal productivity of labor and/or when there is sufficient additional hiring in complementary sectors. Wages are unlikely to rise when workers cannot push for their share of productivity...	Daron Acemoğlu, Simon Johnson
8	2022	Automation and Low-Skill Labor ↗ This paper provides a foundational task-based model directly addressing how automation affects wages and inequality across skill levels, which aligns with the project's core themes of distributional effects and skill complementarity. Although it focuses on general automation rather than specifically on recent AI advancements, its theoretical framework is highly relevant for analyzing the causal mechanisms of labor market displacement and wage impacts. We present a task-based model in which high- and low-skill workers compete against machines in the production of tasks. Low-skill (high-skill) automation corresponds to tasks performed by low-skill (high-skill) labor being taken over by capital. Automation displaces the type of labor it directly affects, depressing its wage. Through ripple effects, automation also affects the real wage of other workers. Counteracting these forces, automation creates a positive productivity effect, pushing up the price of all factors. Because capital adjusts to keep the interest rate constant, the productivity effect dominates in the long run. Finally, low-skill (high-skill) automation increases (reduces)...	Katja Mann, Dario Pozzoli
8	2024	The impact of artificial intelligence in the early retirement decision ↗ This paper directly addresses the project's core themes of AI exposure measurement and distributional effects by analyzing how AI impacts early retirement decisions across different skill levels and occupational contexts. It provides valuable empirical evidence on the causal effects of AI on labor market transitions, specifically highlighting the interaction between worker education and occupation-specific AI characteristics. Abstract This paper examines the impact of Artificial Intelligence (AI) on early retirement (ER) decisions in Europe. For the analysis, we utilize microdata from the Survey of Health, Ageing and Retirement in Europe, along with occupation-level data on AI advances and AI exposure. Initially, we investigate the influence of AI advances and AI exposure separately, finding in both instances a significant reduction in ER likelihood, though this only applies to workers with higher education. Subsequently, we explore the interaction between AI advances and AI exposure concerning ER probability. This interaction proves critical in determining AI’s impact on ER transitions. Specifically, we observe...	Pablo Casas, Concepción Román
8	2025	Generative AI may create a socioeconomic tipping point through labour displacement ↗ This paper directly addresses the project's core questions regarding the aggregate labor market effects of AI, specifically focusing on large-scale labor displacement and its macroeconomic consequences. It provides relevant empirical modeling and policy implications concerning the decoupling of productivity from human input, which aligns with the research theme of economywide employment and wage effects. Work is fundamental to societal prosperity and mental health, providing financial security, a sense of identity and purpose, and social integration. Job insecurity, underemployment and unemployment are well-documented risk factors for mental health issues and suicide. The emergence of generative artificial intelligence (AI) has catalysed debate on job displacement and its corollary impacts on individual and social wellbeing. Some argue that many new jobs and industries will emerge to offset the displacement, while others foresee a widespread decoupling of economic productivity from human input threatening jobs on an unprecedented scale. This study explores the conditions under which both...	Jo‐An Occhipinti, William Hynes, Ante Prodan, et al.
8	2025	AI and policy: what makes AI different? ↗ The paper directly addresses the project's core themes by discussing AI's differential impacts on labor markets, specifically highlighting displacement for low-skill workers and augmentation for high-skill workers, which aligns with the inquiry into winners, losers, and skill complementarity. It also covers aggregate effects on productivity and inequality, providing relevant macroeconomic context and policy implications that are central to the researcher's investigation of AI's economic reshaping. Artificial intelligence (AI) is no longer a speculative or emerging technology but a transformative force affecting firms, workers and economic institutions. Recent studies show its deep integration into economies, highlighting its dual role as a driver of innovation and a source of significant social changes both at the macro and micro levels. In Europe, employment shares in AI-exposed occupations have increased, particularly among younger and more skilled workers, as shown by Albanesi et al. (2025). These changes, however, vary across EU countries, depending on education systems, regulatory frameworks and technology diffusion rates. In the United States, research by Bonfiglioli et al...	Emilio Calvano, Giacomo Calzolari
8	2024	Does increasing robot density exacerbate wealth inequality? ↗ The paper closely aligns with the project's focus on distributional effects and inequality by examining how automation impacts household wealth, specifically for low-skill and young workers. It provides valuable empirical evidence on the mechanisms of skill complementarity and substitution within a task-based framework, although it focuses on industrial robots rather than AI specifically. This article expands the economic consequences of applying automation technology beyond the labor market to encompass wealth distribution. It empirically investigates the effects of changes in robot density on household wealth inequality and potential mechanisms. By using three-digit industry codes provided by the China Census 1 % sampling data in 2015, this paper achieves a more accurate matching of industrial robot data with individual data and employs instrumental variables to alleviate potential endogeneity bias. This paper finds that increasing robot density exacerbates the inequality of family wealth, and this effect has a particularly significant impact on young labor force and...	Nan Jia, Xueling Yan, Shule Yu
8	2023	AI exposure predicts unemployment risk ↗ This paper directly addresses the project's core theme of measuring AI exposure by empirically testing the predictive power of various exposure models on actual unemployment outcomes. It contributes to the causal effects and distributional effects questions by demonstrating that ensemble methods outperform individual models in predicting job separations. Is artificial intelligence (AI) disrupting jobs and creating unemployment? Despite many attempts to quantify occupations' exposure to AI, inconsistent validation obfuscates the relative benefits of each approach. A lack of disaggregated labor outcome data, including unemployment data, further exacerbates the issue. Here, we assess which models of AI exposure predict job separations and unemployment risk using new occupation-level unemployment data by occupation from each US state's unemployment insurance office spanning 2010 through 2020. Although these AI exposure scores have been used by governments and industry, we find that individual AI exposure models are not predictive of...	Morgan R. Frank, Yong‐Yeol Ahn, Esteban Moro
8	2024	THE IMPACT OF AI ON US LABOR MARKETS ↗ This paper directly addresses the project's core theme of whether AI augments or substitutes for labor by analyzing the dual impacts of job displacement and creation in the US labor market. It provides empirical evidence on the relationship between AI models and human tasks, which is central to understanding task reorganization and the causal effects of AI adoption on employment and productivity. This research explores the consequences of AI integration in the labor market. As AI shapes various industries, it brings a dual impact: displacing some jobs while creating others. The automation driven by AI could be a threat to routine tasks, potentially leading to the displacement of specific roles within the routine tasks. However, AI also creates new job opportunities, particularly in AI development and related fields. This study aims to analyze the multifaceted impact of AI on US jobs, considering displacement, creation, and skills. The research considered the following aspects: Evaluation of job displacement and creation, skill shifts, the quality of AI’s impact on performance, and...	Austine Unuriode, Stanley Okoro, Osariemen Afolabi, et al.
8	2025	Impact Of Artificial Intelligence On Employment And Workforce Development: Risks, Opportunities, And Socioeconomic Implications ↗ [Title only] The title directly addresses the core theme of AI's impact on employment, aligning with the project's investigation into causal effects on employment and inequality. It also covers workforce development and socioeconomic implications, which are relevant to understanding distributional effects and the future of work. No abstract available.	Dimple Patil
8	2023	Robot Adoption and Employment Adjustment: Firm-Level Evidence from China ↗ This paper provides direct empirical evidence on how automation technologies affect firm-level employment, specifically focusing on productivity gains and the demand for high-skilled labor. It closely aligns with the project's core themes of labor market effects, skill complementarity versus substitution, and firm reorganization in response to technological adoption. This paper investigates the effect of robot adoption on employment adjustment at the firm level in China. Using an instrument that exploits the source of the firm-level variation in robot adoption, we find that using industrial robots increases employment within firms, especially when focusing on the employment of high-educated and high-skilled individuals. We further analyze the mechanisms underlying the effect of robots and find strong support for the productivity effect that robot adopters gain productivity improvement and thereby increase their demand for labor. In addition, we find limited evidence for the intra-industry labor reallocation to robot-adopting firms from their...	Ting Wang, Yi Zhang, Chun Liu
8	2025	Organizational Technology Ladders: Remote Work and Generative AI Adoption ↗ [Title only] This title directly addresses the project's core themes of how firms reorganize work and the specific impact of generative AI on organizational structures. It likely explores the interaction between remote work and AI adoption, which is highly relevant to understanding task reorganization and the changing nature of work within firms. No abstract available.	Gregor Schubert
8	2024	Evaluation of Task Specific Productivity Improvements Using a Generative Artificial Intelligence Personal Assistant Tool ↗ This paper directly addresses the project's core theme of generative AI productivity experiments by empirically measuring task-specific performance improvements. It provides relevant evidence on how AI tools augment labor in specific office tasks, although the context is limited to a single firm rather than broader labor market dynamics. This study evaluates the productivity improvements achieved using a generative artificial intelligence personal assistant tool (PAT) developed by Trane Technologies. The PAT, based on OpenAI’s GPT 3.5 model, was deployed on Microsoft Azure to ensure secure access and protection of intellectual property. To assess the tool’s productivity effectiveness, an experiment was conducted comparing the completion times and content quality of four common office tasks: writing an email, summarizing an article, creating instructions for a simple task, and preparing a presentation outline. Sixty-three (63) participants were randomly divided into a test group using the PAT and a control group performing...	Brian S. Freeman, Kendall Arriola, Dan Cottell, et al.
8	2024	Generative AI, Adoption and the Structure of Tasks ↗ [Title only] This title directly addresses core themes of AI exposure measurement and task-based frameworks by explicitly linking generative AI to the structure of tasks. It suggests a high likelihood of discussing how AI adoption alters task composition, which is central to understanding augmentation versus substitution effects in labor markets. No abstract available.	Laurence Ales, Christophe Combemale, Krishnan Ramayya
8	2023	Spatial distribution and characteristics of vulnerable occupations to artificial intelligence: cases from South Korea ↗ This paper directly addresses the core theme of measuring AI exposure across occupations by analyzing task vulnerability using patent data and occupational classifications in South Korea. It provides empirical evidence on which workers are most at risk, highlighting a shift from unskilled to middle- and high-skilled roles, which aligns with the project's investigation into distributional effects and task-based frameworks. With the advent of the Fourth Industrial Revolution, Artificial Intelligence (AI) has become increasingly prevalent and is expected to replace a range of human tasks. In this context, this study seeks to identify occupations that are vulnerable to AI, and focuses on their occupational and spatial characteristics. Korean patent data within Google Patents and tasks of occupations based on ISCO-08 were collected and analyzed via dependency parsing to reveal the corresponding tasks and occupations with AI's technical characteristics. Our analysis highlights the vulnerability of tasks that have remained unchanged for a long time, while managerial professions and other occupations involving new...	Seongjun Yang, Donghyun Kim
8	2021	Exposure of occupations to technologies of the fourth industrial revolution ↗ This paper closely relates to the project's core theme of measuring AI and 4IR exposure across occupations, offering a novel patent-based indicator that distinguishes between traditional and fourth industrial revolution technologies. It provides valuable empirical context for understanding how actual technology diffusion impacts job growth, directly addressing the project's focus on exposure measurement and distributional effects. The fourth industrial revolution (4IR) is likely to have a substantial impact on the economy. Companies need to build up capabilities to implement new technologies, and automation may make some occupations obsolete. However, where, when, and how the change will happen remain to be determined. Robust empirical indicators of technological progress linked to occupations can help to illuminate this change. With this aim, we provide such an indicator based on patent data. Using natural language processing, we calculate patent exposure scores for more than 900 occupations, which represent the technological progress related to them. To provide a lens on the impact of the 4IR, we differentiate...	Benjamin Meindl, Morgan R. Frank, Joana Mendonça
8	2025	Routine-Biased Technological Change and Endogenous Skill Investments ↗ This paper closely relates to the project's themes of labor market reshaping by technology and distributional effects, specifically focusing on how workers adapt their skill levels in response to automation. It provides valuable context on endogenous skill investment and the long-term earnings impacts of technological displacement, which are central to understanding AI's effect on inequality and career trajectories. We investigate how individuals alter their educational investments in response to routine-biased technology. We find that individuals growing up in robot-impacted areas are more likely to complete a bachelor’s degree and experience a relative increase in earnings. Changes in the skill premium and opportunity cost appear to drive these effects. To interpret these findings, we estimate a model of endogenous skill acquisition where changes in the demand and supply of skills shape the path of earnings. Counterfactual simulations suggest that the endogenous skill response cannot fully undo the adverse earnings effects of automation unless there are sufficiently generous educational subsidies...	Danyelle Branco, Bladimir Carrillo, Wilman J. Iglesias
8	2023	Automation in an Open, Catching-up Economy: Aggregate and Microeconometric Evidence ↗ This paper closely relates to the project by analyzing how automation reshapes labor markets through changes in task-level labor shares and firm productivity. It provides valuable empirical evidence on whether automation substitutes for or complements labor, directly addressing the core theme of AI skill complementarity versus substitution. Using the universe of firms in Estonia, we study the implications of imports-led and FDI- facilitated automation for productivity and factor shares of tasks and value-added. First, in contrast to the findings for developed economies, we find that the aggregate labour share of value-added for automation-adopting firms is higher than that for non-adopters and has grown, among others, through the reallocation of economic activities towards adopting firms. Second, the aggregate total factor productivity of the adopters concurrently grew faster than that of the non-adopters. Third, from the micro-level study, we find that the estimated labour share of tasks has declined over time among the...	Amaresh Kumar Tiwari
8	2023	Knowledge Trap: Human Experts Distracted by Details When Teaming with AI ↗ [Title only] This paper investigates how human experts interact with AI, specifically focusing on whether AI tools augment or substitute for labor in complex, expert-level tasks. It directly addresses the project's core theme of task-based framework and the distributional effects of AI on specific worker types, such as high-skill experts. No abstract available.	Weiguang Wang, Xiaopei Liu, Xi Zhang, et al.
8	2025	Corporate AI play and short term skill-biased AI change ↗ This paper directly addresses the project's core questions regarding AI skill complementarity versus substitution and the distributional effects on workers by differentiating between automation and augmentation strategies. It provides a task-based theoretical framework and empirical evidence on how corporate AI adoption reshapes labor demand across skill levels, aligning closely with the project's focus on labor market impacts and inequality. We develop a task-based model that illustrates how skill-bias emerges in firms, either because AI competes with lower skills and/or augments more complex jobs, while the extent of bias depends both on the corporate focus on efficiency/innovation and on AI performance scope across the whole range of firm tasks. Based on those predictions, we build an empirical model of skill change across 12 categories to assess whether, and how large, short term changes in skill labor demand correlates with firms use of AI technologies in their business, in the context of AI before genAI development. While AI is skill-biased in favor of more advanced skills, the effect of AI on skill demand is usually...	Jacques Bughin
8	2024	The Macroeconomic Effects of AI Innovation ↗ [Title only] The title directly addresses the core theme of aggregate labor market effects and macroeconomics of AI, which are central to the researcher's project. However, without an abstract, it is unclear if the paper focuses specifically on labor outcomes or broader economic metrics like GDP, introducing some uncertainty about its specific relevance to worker-level impacts. No abstract available.	Andrea Giovanni Gazzani, Filippo Natoli
8	2024	The Role of Human Capital for AI Adoption: Evidence from French Firms ↗ [Title only] This paper directly addresses the project's theme of firm AI adoption by investigating how human capital influences this process within French firms. It provides relevant insights into the skill complementarity dynamics and organizational responses to AI, which are central to understanding the distributional effects and labor market reshaping discussed in the research project. No abstract available.	Luca Fontanelli, Flavio Calvino, Chiara Criscuolo, et al.
8	2025	The Labor Market Impact of Artificial Intelligence: Evidence from US Regions ↗ [Title only] The title suggests a direct examination of AI's effect on labor markets using regional data, which aligns well with the project's interest in causal effects and aggregate evidence. The use of US regions may help address spatial distributional effects, though the specific focus on geographic rather than occupational or task-based metrics requires careful evaluation against the core themes. No abstract available.	Yueling Huang
8	2024	ARTIFICIAL INTELLIGENCE AND SERVICE, INDUSTRIAL, AND AGRICULTURAL EMPLOYMENT: COMPREHENSIVE INTERNATIONAL MACROECONOMIC EVIDENCE ↗ This paper directly addresses the macroeconomic effects of AI on aggregate and sectoral employment, a core theme of the project. It provides relevant international evidence on how AI adoption differentially impacts service, industrial, and agricultural sectors, contributing to the understanding of distributional effects and economywide employment trends. Recent advancements in artificial intelligence (AI) technology have revived concerns about technological unemployment. Regarding the issue, this study examines the impact of AI on employment rates across 17 leading AI countries from 1998 to 2017 using two panel econometric techniques, DOLS and FMOLS, to ensure robust results. For the first time, as far as is known, the effect of AI on employment in distinct sectors is analyzed separately. By uniquely combining different countries and sectors within a macroeconomic framework, this study provides a more comprehensive understanding through a total of eight estimates. The findings indicate that, according to both DOLS and FMOLS techniques...	Yahya Algül
8	2025	Extending “GPTs Are GPTs” to Firms ↗ This paper directly addresses the project's core theme of measuring AI exposure by extending occupational metrics to the firm level, a key component of understanding how firms reorganize work. It provides empirical evidence on how firm characteristics influence LLM exposure, offering relevant insights into the distributional effects and potential productivity gains of AI adoption. We extend Eloundou et al. (2024) to build firm-level measures of exposure to large language models (LLMs) with data from two sources: Eloundou et al. (2024) for occupation-level measures of LLM exposure and Revelio Labs for firm-level employee counts by occupation. The results indicate that companies with more technology workers and AI-skilled employees tend to have higher levels of LLM exposure. We also find that differences in LLM exposure are greater between exposure categories than within them, suggesting that integrating LLMs into corporate systems may lead to significant productivity gains.	Benjamin Labaschin, Tyna Eloundou, Sam Manning, et al.
8	2022	Artificial Intelligence, Robots and Unemployment: Evidence from OECD Countries ↗ This paper is closely related as it directly addresses the causal effects of AI and robotics on aggregate employment and wage inequality using OECD panel data. It provides valuable empirical evidence on distributional impacts by education and age, aligning with the project's focus on who the winners and losers are in the context of technological adoption. Investigating the relation between artificial intelligence, robots and unemployment on a panel of 33 OECD countries covering the 2005—2017 period, we find that a 10% increase in the stock of industrial robots is associated with a 0.42 point increase in the unemployment rate. For artificial intelligence (AI), we use patents as a proxy of AI-related technological capabilities and find a positive correlation with the aggregated unemployment rate, albeit statistically weaker than the one found for robots. We then run the regressions on unemployment rates differentiated by education and age, and observe highly heterogeneous effects between groups. For example, the effect of robots is 2.5 times...	Florent Bordot
8	2025	From automation technology to generative AI: skill heterogeneity in technology’s impact on laborers ↗ This paper directly addresses the project's core themes by analyzing skill heterogeneity and the differential impacts of generative AI versus traditional automation on laborers. It specifically examines distributional effects by skill level, age, and gender, providing relevant empirical evidence on who the winners and losers are in the AI transition. Abstract This is a study of skill heterogeneity in technology’s impact on laborers transitioning from physical automation to cognitive automation. The findings indicate that cognitive skills are a crucial determinant of the extent of technological influence. Considering both technological substitution and technological control, high-skilled and low-skilled workers experience limited technological substitution. However, high-skilled workers are not significantly affected by technological control, whereas low-skilled workers are subjected to stronger technological control. A comparison between automation technology and large language models (LLMs) reveals that the former primarily affects the...	Y.-Z. Zhang
8	2026	Gaps in large language model awareness, usage, and perceptions in the United States: Evidence from a nationally representative longitudinal survey ↗ This paper directly addresses the project's core theme of AI exposure measurement by providing empirical evidence on how LLM adoption varies across worker demographics and occupations. It offers crucial context for understanding distributional effects and the digital divide, which are central to analyzing who benefits from or is displaced by AI technologies. Large language models (LLMs) have the potential to benefit users in both their work and personal lives, but which groups are quickest to adopt them? To investigate awareness, usage, and perceptions of LLMs among US adults across socio-demographic groups-and to track changes over time-we administered a two-wave survey using a nationally representative, probability-based online panel of 12,224 US residents. Across two survey waves spanning 1 year, we observed marked gaps in usage: groups more likely to use LLMs included men, younger adults, those with college education and higher incomes, individuals in more analytical occupations (e.g. STEM), Democratic-leaning respondents, and those with...	Marco Angrisani, Maria Casanova, Nathanael J. Fast, et al.
8	2025	Agentic AI and the Future of Work: Transforming Labor Markets, Economic Structures, and Workforce Development ↗ [Title only] The title directly addresses the transformation of labor markets and workforce development by Agentic AI, which aligns with the project's focus on AI's causal effects on employment and skill structures. It also implicitly covers the themes of task reorganization and aggregate economic impacts, though it lacks specific mention of measurement or distributional inequalities. No abstract available.	Nixalkumar Patel, Heta Chauhan
8	2024	The Artificial Intelligence Premium ↗ [Title only] The title directly suggests an analysis of wage differentials associated with AI skills or adoption, which is central to the project's inquiry into distributional effects and the AI skill premium. It likely provides empirical evidence on whether AI complements specific labor types, addressing core questions about winners and losers in the labor market. No abstract available.	Runfeng Yang, Yufeng Mao, Massimiliano Caporin, et al.
8	2024	De-Routinization in the Fourth Industrial Revolution - Firm-Level Evidence ↗ The paper directly addresses the project's theme of task reorganization and firm-level responses to technological adoption by providing evidence on how frontier technologies drive de-routinization. It offers relevant insights into the distributional effects of technology on employment types, distinguishing between large adopters replacing routine jobs and others experiencing growth, which aligns with the project's focus on winners, losers, and firm reorganization. This paper examines the extent to which aggregate-level de-routinization can be attributed to firm-level technology adoption during the most recent technological expansion. We use administrative data and a novel firm survey to distinguish frontier technologies from older technologies. We find that adopters of frontier technologies contribute substantially to deroutinization. However, this is driven only by a subset of these firms: large adopters replace routine jobs and less routine-intensive adopters experience faster growth. These scale and composition effects reflect firms' readiness to adopt and implement frontier technologies. Our results suggest that an acceleration of technology...	Melanie Arntz, Sabrina Genz, Terry Gregory, et al.
8	2021	How Do Workers Adjust When Firms Adopt New Technologies? ↗ This paper directly addresses the project's core theme of how firms reorganize work and adjust to AI and digital technology adoption by examining causal effects on employment stability and wages. It provides crucial empirical evidence on distributional outcomes, identifying which worker groups benefit or struggle during technological transitions. IZA DP No. 14626 AUGUST 2021 How Do Workers Adjust When Firms Adopt New Technologies? We investigate how workers adjust to firms’ investments into new digital technologies, including artificial intelligence, augmented reality, or 3D printing. For this, we collected novel data that links survey information on firms’ technology adoption to administrative social security data. We then compare individual outcomes between workers employed at technology adopters relative to non-adopters. Depending on the type of technology, we find evidence for improved employment stability, higher wage growth, and increased cumulative earnings in response to digital technology adoption. These beneficial...	Sabrina Genz, Terry Gregory, Markus Janser, et al.
8	2022	Computerization of White Collar Jobs ↗ This paper closely aligns with the project's focus on AI as a form of technological adoption by examining its impact on white-collar wages, employment, and task reorganization. It provides empirical evidence on skill complementarity versus substitution and distributional effects, which are central themes in the researcher's inquiry into how technology reshapes labor markets. We investigate the impact of computerization of white-collar jobs on wages and employment. Using online job postings from 2007 and 2010--2016 for office and administrative support (OAS) jobs, we show that when firms adopt new software at the job-title level they increase the skills required of job applicants. Furthermore, firms change the task content of such jobs, broadening them to include tasks associated with higher-skill office functions. We aggregate these patterns to the local labor-market level, instrumenting for local technology adoption with national measures. We find that a 1 standard deviation increase in OAS technology usage reduces employment in OAS occupations by about 1...	Marcus Dillender, Eliza Forsythe
8	2025	Transforming the Working Style of Call Center Agents Through Generative AI ↗ This paper directly examines how generative AI augments labor by transforming the workflow and productivity of call center agents, a core theme of the project. It provides empirical evidence on task reorganization and performance metrics, offering specific insights into the intersection of AI adoption and worker output in a defined occupational setting. As Generative Artificial Intelligence (Gen AI) is evolving rapidly, there is a significant change in the approach by the contact center industry with respect to work culture. Historically, customer service agents working in a contact center used to depend significantly on static scripts and fragmented information systems, which thereby resulted in delayed resolutions, cognitive overload, and made them deliver inconsistent customer experiences. This study explores the paradigm shift that's occurring in contact service centers through implementing Gen AI. Real-time intent recognition, contextual response generation, and personalized engagement across channels are some novel capabilities...	Satya Karteek Gudipati
8	2025	Human-AI Collaboration in Software Development: A Mixed-Methods Study of Developers’ Use of GitHub Copilot and ChatGPT ↗ This paper directly addresses the project's theme of firm-level AI adoption and task reorganization by examining how software developers integrate Generative AI tools into their workflows. It provides relevant empirical insights into the mechanisms of human-AI collaboration and the organizational factors influencing productivity and adoption in a key technical occupation. The integration of AI-powered coding assistants into software development presents both opportunities and challenges. These tools promise enhanced productivity and code quality while introducing significant concerns related to workflow integration, accuracy, reliability, and ethical considerations. This study investigates how software developers interact with Generative AI (GenAI) tools in a large-scale public sector organization. Using a mixed-methods approach, we analyze 13 interviews and a survey with 114 respondents to explore the impact of GitHub Copilot (IDE-integrated AI) and ChatGPT (conversational AI assistant) on software engineering workflows. The Human-AI Collaboration and...	Viktoria Stray, Astri Barbala, Viggo Tellefsen Wivestad
8	2024	Machine Knowledge Capital and the Future of Work ↗ [Title only] The title explicitly links 'machine knowledge' to 'the future of work,' suggesting a direct investigation into how AI capital substitutes or complements labor. This aligns with the project's core themes of AI skill complementarity, aggregate labor market effects, and the redistribution of value in the age of automation. No abstract available.	Dan Ciuriak
8	2024	AI Adoption and Firm Demand for Workers and Skills: Insights from Online Job Postings ↗ This paper directly addresses the project's core question of whether AI augments or substitutes for labor by analyzing firm-level hiring trends in response to AI adoption. It provides relevant empirical evidence on how AI exposure affects demand for specific skills and occupations, particularly highlighting the transition of roles and the augmentation of high-skilled workers. The latest Artificial Intelligence (AI) tools can perform some of the complex tasks that highly skilled and well-paid workers perform. To investigate their effects on demand for workers and skills, we compared hiring trends in Australian firms that were adopting AI and those that were not. Job postings grew significantly faster in firms that had adopted AI, even after controlling for firm size, geography and industry. This accelerated growth in job postings included occupations that were highly exposed to AI. The number of skills sought in job postings was also growing faster for AI exposed occupations, especially if they were being recruited by AI adopting firms. Some formerly non-AI...	Claire Mason, Haohui Chen, David Evans
8	2024	Внедрение генеративного ИИ в деятельность финансовой компании: ожидания, эффективность, обучение персонала ↗ This paper directly addresses the project's theme of generative AI productivity experiments by providing empirical evidence on the quantitative economic impact of AI training on employee productivity within a financial firm. It aligns closely with the core questions regarding how AI tools augment labor and how firms reorganize work and measure efficiency gains from AI adoption. Введение. Компании реального сектора активно ищут новые технологии, которые могут минимизировать рутинные задачи и поддерживать процесс принятия управленческих решений. Генеративный искусственный интеллект (ГИИ) становится перспективным инструментом для решения этих задач и получения количественного экономического эффекта. Тем больший интерес представляют компании реального сектора, которые уже внедрили эти технологии и получили ощутимый экономический эффект. Цель исследования — оценить количественный экономический эффект от обучения сотрудников финансовой компании использованию технологий ГИИ и их интеграции в профессиональную деятельность. Методы. В ходе исследования анализировались...	С.В. Кобелев, А.А. Морозкин, Ekaterina Alexeevna Kharitonova
8	2024	The impact of artificial intelligence on creative industries: Freelancers’ anxieties and concerns ↗ This paper directly addresses the project's focus on distributional effects and task reorganization within creative occupations, a key sector for generative AI. It provides empirical evidence on worker anxiety, adoption behaviors, and perceived displacement risks, which are central to understanding how AI reshapes labor markets and affects entry-level or gig workers. The article examines the impact of the rapid development of artificial intelligence (AI) technologies on the creative industries and the concerns of workers in this field regarding the potential deterioration of their working conditions and displacement from the labor market. The aim of the study is to identify the degree of concern among freelancers engaged in intellectual and creative professions regarding competition with AI and to assess their perception of AI’s current capabilities in making creative content. The empirical basis was provided by online survey data of 778 Russian freelancers receiving jobs through the Freelance.ru digital platform, conducted in spring 2024. It was found...	Denis Strebkov
8	2023	Machines and Superstars: Technological Change and Top Labor Incomes ↗ [Title only] This paper likely investigates how technological advancements contribute to rising wage inequality and top income shares, directly addressing the project's interest in distributional effects and AI's impact on labor markets. Although the specific focus on artificial intelligence versus general technological change is ambiguous without the abstract, the core mechanism of technology favoring high-skill 'superstars' aligns strongly with the project's themes on inequality and skill complementarity. No abstract available.	Donghyun Suh
8	2023	"This Time it's Different" - Generative Artificial Intelligence and Occupational Choice ↗ [Title only] The title directly addresses the project's core interest in how AI exposure reshapes labor markets, specifically focusing on the critical dimension of occupational choice and potential shifts in worker allocation. It likely contributes to understanding distributional effects and task reorganization by examining whether generative AI drives structural changes in which occupations workers enter or leave. No abstract available.	Daniel Göller, Christian Gschwendt, Stefan C. Wolter
8	2025	AI for Fairness? The Role of Generative Tools in Shaping Freelancer Success ↗ [Title only] This paper directly addresses the distributional effects of AI on workers, specifically focusing on freelancers as a key segment of the online labor market mentioned in the core themes. It investigates whether generative tools act as augmenting or substituting forces and how they impact worker success, thereby shedding light on inequality and the winner-loser dynamics of AI adoption. No abstract available.	Jingman Cao, Zhonglin Li
8	2025	Automation, AI, and the Intergenerational Transmission of Knowledge ↗ [Title only] This title directly addresses the distributional effects of AI, specifically focusing on how it impacts intergenerational knowledge transfer, which aligns with the project's interest in career ladders and skill transmission. It likely explores whether AI substitutes for or augments experienced workers' ability to train younger entrants, a key mechanism in understanding entry-level labor market dynamics. No abstract available.	Enrique Ide
8	2023	Artificial Intelligence and the Labor Force: A Data-Driven Approach to Identifying Exposed Occupations ↗ This paper directly addresses the project's core theme of measuring AI exposure across occupations by developing a data-driven methodology using NLP to quantify exposure to specific AI technologies. It provides essential empirical context for understanding how different types of AI impact labor market outcomes, aligning closely with the investigation of task-based frameworks and distributional effects. The authors explore the relationship between occupational exposure and technologies, wages, and employment related to artificial intelligence (AI). Using natural language processing (NLP), the authors evaluate occupational exposure to all technology patents in the United States, as well as to specific AI technologies, including machine learning, NLP, speech recognition, planning control, AI hardware, computer vision, and evolutionary computation.	Tobias Sytsma, Sousa, Éder M.
8	2025	Generative AI in Equilibrium: Evidence from a Creative Goods Marketplace ↗ This paper closely relates to the project by examining how generative AI reshapes labor markets, specifically focusing on task substitution and firm/work reorganization in a creative sector. It provides empirical evidence on the causal effects of AI adoption on worker participation, productivity (via production increases), and entry/exit dynamics, which are central to the project's inquiry into winners and losers and the augment vs. substitute debate. We study how generative artificial intelligence (GenAI) affects creative goods markets using data from a large stock images marketplace. In December 2022, the platform announced it would allow artists to sell GenAI-produced images, subject to two conditions: all GenAI images must be labeled, and GenAI would be prohibited in certain markets. We exploit this policy variation using a difference-in-differences design. We estimate a 136\% increase in image production and 47% increase in active artists, accompanied by a 15% decline in non-GenAI production and 29% decline in non-GenAI active artists. On the demand side, total sales increase by 82%, but non-GenAI sales fall by 28% and non-GenAI...	Samuel Goldberg, H.K. Lam
8	2025	Field Choice, Skill Specificity, and Labor Market Disruptions ↗ This paper is closely related as it explicitly models the wide-scale adoption of AI and its aggregate and distributional economic impacts. It contributes directly to the project's themes by examining how skill specificity and field choice influence labor market adjustments to technological shocks, offering a relevant framework for understanding long-term workforce responses. We argue that college students' field-of-study choices significantly influence how economies respond to labor market disruptions.To do so, we develop and estimate a framework featuring forward-looking students who choose a field of study when entering college, and subsequently make decisions over occupations after graduating and entering the labor market.Different fields endow workers with distinct comparative advantages and varying costs associated with switching occupations.Simulating both a trade war and wide scale adoption of AI, we use our model to make three points.First, relative to models that ignore how new cohorts adjust their field-of-study choices, our framework predicts larger...	Valérie Smeets, Lin Tian, Sharon Traiberman
8	2025	The impact of China’s National Artificial Intelligence Innovation Pilot Zones on skill demand: a place-based policy perspective ↗ [Title only] This paper directly addresses the core theme of how AI reshapes labor markets by examining the causal impact of a major policy intervention on skill demand. It provides relevant place-based evidence on AI exposure and its distributional effects across different worker types. No abstract available.	S. Sun
8	2026	Automation as an Equalizer: How Easy‐to‐Use Technologies Narrow Skill Gaps Between Low‐ and High‐Skilled Workers ↗ This paper directly addresses the project's core theme of whether AI tools augment or substitute for labor, specifically focusing on skill complementarity. It provides relevant evidence on how generative AI and LLMs narrow skill gaps, contributing to the understanding of distributional effects and worker productivity changes. ABSTRACT This paper investigates the labor market effects of the two most disruptive technologies of the past decade–industrial robots and artificial intelligence (AI). By reviewing the empirical literature and discussing existing models, we explore how these technologies affect workers based on their level of skills. The reviewed studies indicate that, contrary to popular belief, AI use does not hurt workers, including the low‐skilled. On the contrary, the ease of using these technologies, particularly the more recent iterations of AI and large language models (LLMs), makes them complimentary to the low‐skilled workforce, enabling them to reach productivity levels closer to those of high...	Shahab Sharfaei, Nopadol Rompho
8	2024	Not feeling like winners: The higher educated and the AI revolution ↗ This paper directly addresses the project's inquiry into how AI exposure varies by skill level and occupation, specifically challenging the notion that only lower-educated workers are at risk by highlighting significant vulnerability among the higher educated. It contributes to the core theme of distributional effects and AI skill complementarity vs. substitution by identifying specific white-collar industries and task types where AI exposure creates perceived technological redundancy. Technological redundancy was so far seen as mainly a concern for lower to medium educated workers, but recent advances in machine learning and algorithmic decision-making technology – i.e., “artificial intelligence” (AI) – have raised concerns that the higher educated could now be affected as well. Adding to an emerging literature on the effects of AI exposure, I show that higher educated workers indeed feel that they are at risk of technological redundancy, and that these perceptions are related to objective exposure to AI technology. Specifically, I combine comparative survey data on workers’ subjective technological vulnerability with indicators of objective industry-level AI exposure...	Carlo Knotz
8	2020	Innovación y empleo. Paradojas sociales y económicas. ↗ This paper directly addresses the core theme of whether AI and automation act as substitutes for or complements to labor, specifically examining the paradoxes of innovation versus employment. It provides relevant theoretical context and critical discussion on the distributional effects of new technologies on labor markets and wealth distribution. La relación entre la innovación y el empleo es un tema debatido desde los inicios del capitalismo. A menudo, domina la visión pesimista, donde la nueva tecnología tiende a desplazar la mano de obra existente, como es el caso de la actual discusión sobre la robotización y la inteligencia artificial. Sin embargo, la evidencia empírica muestra que las tasas de empleo siguen creciendo a pesar de la introducción de nuevas máquinas. El objetivo del artículo es revisar las tendencias y teorías respecto a la relación innovación versus empleo tanto en términos cuantitativos como cualitativos. Además, se discute de forma crítica el impacto que tienen las nuevas tecnologías sobre las relaciones...	Joost Heijs, Guillermo Arenas
8	2022	Productivity and Wages: What Was the Productivity-Wage Link in the Digital Revolution of the Past, and What Might Occur in the AI Revolution of the Future? ↗ This paper directly addresses the project's core interest in the distributional effects of AI on wages and inequality by comparing current AI trends to historical digital revolutions. It provides relevant empirical context on skill-biased technological change and its impact on the wage-productivity link, which aligns with the research focus on winners and losers across skill levels. Wages have been spreading out across workers over time -or in other words, the 90th/50th wage ratio has risen over time. A key question is, has the productivity distribution also spread out across worker skill levels over time? Using our calculations of productivity by skill level for the U.S., we show that the distributions of both wages and productivity have spread out over time, as the right tail lengthens for both. We add OECD countries, showing that the wage-productivity correlation exists, such that gains in aggregate productivity, or GDP per person, have resulted in higher wages for workers at the top and bottom of the wage distribution. However, across countries, those workers in...	Edward P. Lazear, Kathryn Shaw, Grant Hayes, et al.
8	2025	The Labor Market Impact of Digital Technologies ↗ [Title only] This title suggests a broad examination of how digital technologies, including AI, affect the labor market, which aligns with the project's interest in causal effects on employment, wages, and inequality. However, the lack of specific mention of machine learning or generative AI in the title introduces uncertainty about whether the paper addresses the modern AI landscape central to the researcher's project. No abstract available.	Sangmin Aum, Yongseok Shin
8	2025	Lower-skilled occupations face greater upskilling pressure in U.S. job ads ↗ This paper directly addresses the project's core theme of measuring AI exposure and task-based skill changes by analyzing reskilling pressures in job ads. It provides specific empirical evidence on how lower-skilled occupations face greater upskilling requirements, which is crucial for understanding distributional effects and labor market reorganization due to new technologies. Substantial scholarship has estimated the susceptibility of jobs to automation, but little has examined how job contents evolve in the information age as new technologies substitute for tasks, shifting required skills rather than eliminating entire jobs. Here we explore patterns of occupational skill change and characterize occupations and workers subject to the greatest re-skilling requirements in the United States. Recent work found that changing skill requirements are greatest for STEM occupations in the 2010s. Nevertheless, analyzing 167 million online job posts covering 721 occupations, we find that when accounting for distance between skills, skill change is greater for lower-skilled...	Di Tong, Lingfei Wu, James A. Evans
8	2025	The Economics of Bicycles for the Mind ↗ This paper provides a formal theoretical model that directly addresses the core themes of AI skill complementarity versus substitution and task reorganization by distinguishing between judgment and implementation skills. It synthesizes empirical literature on AI's impact on productivity and inequality, offering a mechanistic framework relevant to understanding how cognitive tools reshape labor markets and decision-making authority within firms. Steve Jobs described computers as "bicycles for the mind," a tool that allowed people to dramatically leverage their capabilities.This paper presents a formal model of cognitive tools and technologies that enhance mental capabilities.We consider agents engaged in iterative task improvement, where cognitive tools are assumed to be substitutes for implementation skills and may or may not be complements to judgment, depending on their type.The ability to recognise opportunities to start or improve a process, which we term opportunity judgment, is shown to always complement cognitive tools.The ability to know which action to take in a given state, which we term payoff judgment, is not...	Ajay Agrawal, Joshua S. Gans, Avi Goldfarb
8	2026	Not a Typical Firm: Capital–Labor Substitution and Firms' Labor Shares ↗ This paper directly addresses the project's core themes of firm reorganization in response to automation and the distributional effects of AI on labor shares. It provides a theoretical framework linking capital-labor substitution to firm heterogeneity, which is highly relevant to understanding aggregate labor market effects and who the winners and losers are. The US labor share has declined, especially in manufacturing and retail. Yet the labor share of a typical firm in these sectors has risen. We introduce a model where firms incur fixed costs to automate tasks. A decline in the price of capital goods used for automation reproduces the observed patterns: large firms automate tasks, reducing the aggregate labor share, while the median firm continues to operate a labor-intensive technology. When calibrating the automation fixed costs to match the observed adoption heterogeneity, the model generates the aggregate and firm-level facts quantitatively in response to lower capital prices, especially in manufacturing. (JEL D21, D33, E25, L60, O32)	Joachim Hubmer, Pascual Restrepo
8	2026	Politically Coded: CEO Partisanship and Firm AI Adoption ↗ This paper directly addresses the firm-level AI adoption dimension of the project by investigating how CEO partisanship causally influences corporate AI hiring strategies. It provides valuable empirical context on the determinants of firm AI adoption, a key mechanism underlying the aggregate labor market effects examined in the research. Using FEC political donation records to measure CEO partisanship and job posting data from Revelio Labs to measure corporate AI adoption, we document a robust partisan gap: firms led by Republican-leaning CEOs devote a smaller share of their hiring to AI than those led by Democratic-leaning CEOs. Instrumental variable estimation exploiting the partisan composition of the local CEO labor market suggests a causal interpretation. Restricting to CEOs appointed before AI became a corporate priority produces similar results. The partisan gap concentrates in high AI-exposure industries and persists even when peers are visibly adopting AI. We also find that appointing a Republican-leaning CEO in a...	Yongqiang Chu, Alicia Li, Ruiyao Zhu
8	2023	New technologies and jobs in Europe ↗ This paper closely relates to the project by providing macro-level evidence on AI exposure effects across European occupations, specifically addressing employment shares and the role of skill levels and age. However, it offers limited insight into the causal mechanisms of wage impacts or firm-level task reorganization, which are central to the researcher's specific questions on productivity and inequality. We examine the link between labour market developments and new technologies such as artificial intelligence (AI) and software in 16 European countries over the period 2011-2019. Using data for occupations at the 3-digit level in Europe, we find that on average employment shares have increased in occupations more exposed to AI. This is particularly the case for occupations with a relatively higher proportion of younger and skilled workers. This evidence is in line with the Skill-Biased Technological Change theory. While there is heterogeneity across countries, very few countries show a decline in the employment shares of occupations more exposed to AI-enabled automation. Country...	Stefania Albanesi, António Dias da Silva, Juan F. Jimeno, et al.
8	2024	This Time It's Different – Generative Artificial Intelligence and Occupational Choice ↗ The paper closely relates to the project by empirically analyzing how generative AI exposure influences occupational choice and search behavior among young workers. It provides key evidence on the causal effects of AI on entry-level labor market dynamics and task-based risk perception, aligning directly with the research themes of AI's impact on career ladders and distributional effects. We show the causal influence of the launch of generative artificial intelligence (AI) in the form of ChatGPT on the search behavior of young people for apprenticeship vacancies. In order to estimate the short- and medium-term effects, we use a variety of methods, including a difference-in-discontinuity approach exploiting the exogenous nature of the unanticipated launch of ChatGPT in 2022. There is a strong short- and medium-term decline in the intensity of searches for vacancies, which suggests great uncertainty among the affected cohort. Occupations with a high proportion of cognitive tasks and with high demands on language skills were particularly affected by the decline. Interestingly...	Daniel Göller, Christian Gschwendt, Stefan C. Wolter
8	2024	IMPACT AND REGULATIONS OF AI ON LABOR MARKETS AND EMPLOYMENT IN USA ↗ [Title only] The title explicitly addresses the impact of AI on US labor markets and employment, aligning directly with the project's core questions on aggregate effects and causal impacts. However, the inclusion of 'Regulations' suggests a potential focus on policy rather than purely economic mechanisms or task-based analysis, which may slightly dilute its relevance to the specific micro-foundational themes of skill complementarity and firm reorganization. No abstract available.	Prabin Adhikari, Prashamsa Hamal, Francis Baidoo
8	2024	Threats or Opportunities? Enhancing Firm Performance in the Era of Generative AI ↗ [Title only] The title directly addresses firm-level outcomes and the dual nature of generative AI, which aligns with the project's core themes on firm reorganization and aggregate effects. Although it lacks specific labor market metrics in the title, firm performance is a strong proxy for the underlying productivity and employment dynamics the researcher seeks to understand. No abstract available.	Seung Jong Lee, Julian Lehmann, Heewon Chae, et al.
8	2024	Artificial intelligence and the changing demand for skills in Canada ↗ This paper directly addresses the core theme of AI skill complementarity and substitution by analyzing how AI exposure shifts labor demand towards management, communication, and social skills in Canada. It provides empirical evidence on task-based changes and worker requirements, offering relevant insights into who the potential winners and losers are in the context of AI adoption. Most workers who will be exposed to artificial intelligence (AI) will not require specialised AI skills (e.g. machine learning, natural language processing, etc.). Even so, AI will change the tasks these workers do, and the skills they require. This report provides first estimates for Canada on the effect of artificial intelligence on the demand for skills in jobs that do not require specialised AI skills. The results show that the skills most demanded in occupations highly exposed to AI are management, communication and digital skills. These include skills in budgeting, accounting, written communication, as well as competencies in basic word processing and spreadsheet software. The results...	Andrew Green
8	2024	Corporate Ai Play and Skill-Biased Ai Change ↗ [Title only] This title directly addresses the core project themes of AI skill complementarity versus substitution and the distributional effects of AI on different worker groups. It likely explores how corporate adoption strategies interact with broader economic shifts, making it highly relevant to understanding the causal effects of AI on labor markets. No abstract available.	Jacques Bughin
8	2024	AI and Productivity: The Impact of ChatGPT's Release on Blogging ↗ This paper directly addresses the core theme of generative AI productivity experiments by providing quasi-experimental evidence on how ChatGPT adoption affects worker output in an online labor market. It specifically examines the nuanced effects of AI on productivity, novelty, and popularity, offering insights into whether AI tools augment or substitute for creative labor tasks. <div> Artificial intelligence (AI) is one of the fastest-growing technologies in history, offering innovative tools that can enhance worker productivity. In this paper, we examine the influence of ChatGPT's release on the output of workers who adopt it in a quasi-experiment. We study the productivity, popularity, and novelty of bloggers and their work on Medium.com. Popularity and novelty reflect productive efforts to engage one's readers and expand the scope of one's work, respectively. We measure AI use with the least biased detection tool available at the time of this study. We use Coarsened Exact Matching to match ChatGPT-adopters with similar bloggers who do not adopt. We then perform...	Joshua Kaisen, Meng Li, Shijie Lu
8	2024	Human in the Center of the Human Algorithmic Collaboration in Analytical and Creative Tasks: Evidence From Quant Finance Competition ↗ [Title only] This paper directly investigates the augmentation versus substitution debate by analyzing human-algorithmic collaboration in quantitative finance, a key domain for understanding task reorganization. It provides empirical evidence on how AI interacts with human workers in analytical tasks, addressing core questions about worker productivity and skill complementarity. No abstract available.	Dequn Teng, Chen Ye, Veronica Martinez
8	2024	Who Benefits from AI? Project-level Evidence on Labor Demand, Operations and Profitability ↗ This paper directly addresses the project's core question of whether AI augments or substitutes for labor by providing causal evidence from a large-scale logistics firm. It offers valuable insights into firm reorganization, the distributional effects on different worker types, and the interplay between hardware and software technologies in reshaping labor demand and profitability. We examine how the adoption of digital automation technology affects labor demand, operations and profitability in the context of the logistics industry. Our data covers 9,300 digital automation projects in a multinational company involving service robots and machine learning-based software from 2019 to 2021, alongside fine-grained labor and operations data. To identify causal effects, we leverage exogenous variation from supply-chain disruptions and travel restrictions during COVID-19 and an import ban on information and communication technologies imposed by the Trump administration. We find that total labor cost increased after the adoption of digital automation technology, attributable...	Erdem Dogukan Yilmaz, Christian Peukert
8	2024	Assessing the Impact of Generative AI on Canadian Labor Market: An Empirical Approach ↗ This paper directly addresses the core theme of measuring AI exposure across occupations by adapting established methodologies to the Canadian context. It provides relevant empirical evidence on which job categories are most susceptible to generative AI, informing the distributional and task-based analysis central to the research project. The rapid advancement and integration of Generative AI and Large Language Models (LLMs) into various sectors raise significant concerns about their impact on the labor market. This research assesses the extent to which occupations in Canada are exposed to these technologies. Using data from the Canadian Occupational and Skills Information System (OaSIS) and adapting the methodology of Felton et al. (2018, 2021, 2023), we calculated AI Occupational Exposure (AIOE) scores for 900 occupations. The findings demonstrate a high correlation between Canadian and U.S. occupations in terms of AI exposure, with Pearson and Spearman coefficients of 0.888 and 0.883, respectively. Approximately 45% of...	Aida Kazemi, Ana Lucia Santos De Araujo, Genesis Andreina Uzcategui Lopez
8	2024	<div> <div> <div> <p><span>What is the Impact of AI on Firms? </span></p> </div> </div></div> ↗ This paper directly addresses the project's core theme of how firms reorganize work and the aggregate effects of AI on productivity and GDP. It provides empirical evidence on AI's complementary impact on worker productivity and early signs of labor displacement, which are central to understanding the causal effects of AI on labor markets. I present a new framework for investigating AI’s impact by analyzing managerial discussions. This forward-looking approach captures firm-specific exposure to AI in near-real time without relying on job postings or resumes. To validate the framework, I show that AI-exposed firms subsequently increase their investments in AI-related labor, suggesting that textual information accurately reflects real resource allocation on average. The framework accommodates a range of topics and can estimate many otherwise hard-to-quantify variables of societal importance. I focus on two foundational agendas: namely, AI’s impact on firms and GDP growth. In the cross-section, AI-exposed firms exhibit higher...	Jiacheng Liu
8	2024	Вплив використання генеративного штучного інтелекту на продуктивність розробників програмних продуктів ↗ This paper directly addresses the project's theme of generative AI productivity experiments by providing empirical evidence that AI tools significantly increase developer productivity and code quality. It offers valuable data on task augmentation in a specific occupation, aligning closely with the investigation of who are the winners and losers in AI-driven labor markets. Означено, що створення та швидке впровадження генеративного штучного інтелекту стає рушійним фактором подальшого розвитку технологічного прогресу. Аналізуючи розвиток більшості галузей, можна стверджувати, що його використання дає змогу отримати додану вартість, оскільки він допомагає частково або повністю автоматизувати низку функцій, які до цього здійснювали наймані працівники. Найбільше застосування штучного інтелекту спостерігається у виробництві, медицині, безпеці, енергетиці. Це свідчить про те, що проходить перерозподіл професій. Тобто виокремлюються окремі з них, які потрібні для роботи зі штучним інтелектом. Насамперед це ті, хто розробляє рішення з його використанням, передусім...	О. Ковалишин, Л. Чухрай, Н. Заплатинський
8	2025	Artificial Intelligence and Technological Unemployment ↗ This paper directly addresses the project's core question regarding the aggregate labor market effects of AI, specifically focusing on technological unemployment and productivity changes. It provides a theoretical framework for how AI adoption impacts employment through labor-search dynamics and wage renegotiation, which is central to understanding the causal effects of AI on workers. How large is the impact of artificial intelligence (AI) on labor productivity and unemployment?This paper introduces a labor-search model of technological unemployment, conceptualizing the generative aspect of AI as a learning-by-using technology.AI capability improves through machine learning from workers and in turn enhances their labor productivity, but eventually displaces workers if wage renegotiation fails.Three distinct equilibria emerge: no AI, some AI with higher unemployment, or unbounded AI with sustained endogenous growth and little impact on employment.By calibrating to the U.S. data, our model predicts more than threefold improvements in productivity in some-AI steady state...	Ping Wang, Tsz-Nga Wong
8	2019	Testing the Automation Revolution Hypothesis ↗ This paper directly addresses the project's core theme of measuring automation exposure by evaluating ONET job features and vulnerability metrics against expert reports. It provides critical empirical context on how traditional task-based measures predict automation levels, which is essential for understanding the measurement and distributional effects of AI in labor markets. Recently, many have predicted an imminent automation revolution, and large resulting job losses. Others have created metrics to predict new patterns in job automation vulnerability. As context to such claims, we test basic theory, two vulnerability metrics, and 251 ONET job features as predictors of 1505 expert reports regarding automation levels in 832 U.S. job types from 1999 to 2019. We find that pay, employment, and vulnerability metrics are predictive (R^2~0.15), but add little to the top 25 O*NET job features, which together predict far better (R^2~0.55). These best predictors seem understandable in terms of traditional kinds of automation, and have not changed over our time period...	Keller Scholl, Robin Hanson
8	2022	Automation and the Changing Nature of Work ↗ [Title only] This title directly addresses the project's core theme of how technological changes, including AI, reshape labor markets and work structures. While it may not explicitly mention modern machine learning, papers with this broad framing typically cover the causal effects of automation on employment, wages, and task reorganization relevant to the researcher's inquiry. No abstract available.	Cecily Josten, Grace Lordan
8	2024	Analyzing Automation Technologies and Their Tasks in Patent Texts Using Natural Language Processing ↗ This paper is closely related as it develops a method to measure AI exposure by categorizing automation tasks from patent texts, directly addressing the project's core theme of AI exposure measurement. It provides valuable empirical context by distinguishing AI's focus on higher cognitive tasks from robots and software, which informs the task-based framework and substitution vs. augmentation debate. Technological advances have changed the labor market, and automation technology is a major factor affecting jobs. To study its impact, it's important to accurately measure automation technologies. Patent text is a novel approach to understanding technological progress and identifying the tasks that different automation technologies can perform. The paper uses a dictionary-based approach to identify automation technologies from patent text and categorizes them into three groups: Robots, Software, and Artificial Intelligence. We find that the number of patents related to these groups has grown exponentially since 1980, with AI patents growing the fastest since the 1990s. Most patents are in...	Tomáš Oleš
8	2023	The economic impact of autonomous technologies and a model for the circular economy ↗ This dissertation is highly relevant as it directly examines the macroeconomic and microeconomic labor market impacts of autonomous technologies, including AI and machine learning, which aligns with the project's core questions on causal effects and distributional outcomes. It provides valuable empirical evidence on how these technologies influence employment, wages, and worker mobility across different demographics and regions. This dissertation offers a comprehensive analysis of the wide-ranging economic implications and labour market consequences of autonomous technologies, while also considering the role of the circular economy in sustaining economic growth. Spanning three thematic areas across nine self-contained essays, the research provides a combination of macroeconomic modelling and microeconomic evidence. The thesis is divided into three main parts, leaving aside the introduction and the concluding remarks. The first part focuses on macroeconomic theoretical modelling of a dual traditional-autonomous economy, employing dynamic general equilibrium models to evaluate the impact of autonomous technologies on...	Pablo Casas
8	2024	Unslicing the Pie: AI Innovation and the Labor Share in European Regions ↗ This paper directly addresses the project's core theme of distributional effects by analyzing how AI innovation impacts the labor share and wage inequality across different skill levels in European regions. It provides valuable empirical evidence on whether AI substitutes for labor, particularly among high- and medium-skill workers, which is central to understanding the macroeconomic and distributive consequences of AI adoption. This paper examines how the development of Artificial Intelligence (AI) affects the distribution of income between capital and labor, and how these shifts contribute to regional income inequality. To investigate this issue, we analyze data from European regions dating back to 2000. We find that for every doubling of regional AI innovation, the labor share declines by 0.6% to 1.6%, potentially reducing it by 0.20 to 0.53 percentage points from an average of 52%, solely due to AI. This new technology has a particularly negative impact on high- and medium-skill workers, primarily through wage compression, while for low-skill workers, employment expansion induced by AI mildly offsets the...	Antonio Minniti, Klaus Prettner, Francesco Venturini
8	2024	Afraid of Automation? Choose Your Training Carefully ↗ [Title only] This title suggests a focus on worker training as a strategic response to automation risks, directly addressing the project's interest in how workers and firms adapt to technological shifts. It likely explores the distributional effects of automation by examining whether specific skill development augments labor or mitigates substitution risks. No abstract available.	Miroslav Štefánik, Lukáš Lafférs, Zuzana Košťálová
8	2025	How Are Technologies Changing the Requirements for Skills and Education? (A Discussion Review) ↗ This review directly addresses the core project theme of AI skill complementarity versus substitution, providing a synthesis of recent literature on how AI reshapes skill demands and wage premiums. It offers valuable context for understanding distributional effects and task reorganization, aligning closely with the project's focus on labor market changes driven by technological progress. The paper explores the changing demand for skills and qualifications under the influence of technological progress and the implementation of artificial intelligence. The research issue is examined in three dimensions: the structure of skills in the labor market, the returns to education, and the demand for specific skills and competences. A systematic analysis of 125 key publications in labor economics and skills research from leading think tanks and academic journals between 2018 and 2023 was conducted to identify the effects that technology has on the labor market and to map the changing structure of skills. The presented analyses of the latest empirical studies allow us to draw...	Vera Maltseva, Mikhail Nikitin
8	2025	Riding the Wave of Artificial Intelligence: A Career Ladder Approach ↗ This paper directly addresses the project's inquiry into who the losers of AI are and how entry-level or easily automatable workers transition to new roles. It provides a task-based analysis of career ladder dynamics, identifying specific occupational shifts from AI-exposed to less exposed jobs, which aligns with the core theme of AI's impact on labor market mobility and skill complementarity. 본 연구는 커리어래더 접근법을 활용하여 인공지능에 의하여 대체가능성이 높은 직업을 확인하고, 이들 직업이 대체가능성이 낮은 직업으로 어떻게 이･전직할 수 있는지를 분석하였다. 분석에 사용된 자료는 ONET과 한국표준직업분류를 매칭하여 구축한 데이터로 직업별 기술과 역량의 중요도 정보를 포함한다. 해당 자료를 기반으로 직업 간 유사도는 유클리디언 거리를 활용하여 측정하였으며, 대체가능성이 높은 직업의 유형 구분은 잠재프로파일분석을 사용하였다. 분석 결과, 대체가능성이 높은 직업을 총 6가지 유형으로 분류하였으며, 직업들은 공통적으로 복잡성과 전문성이 상대적으로 낮은 특징을 보이는 것으로 나타났다. 한편, 이들 직업은 다양한 대체가능성이 낮은 직업으로 이･전직이 가능한데, 대표적으로 자동차 정비원, 채굴, 토목, 화학 관련 직업 등이 있는데, 이러한 대체가능성이 낮은 직업들은 기술적 숙련도뿐 아니라 상황적응 능력, 인지적 역량, 신체적 숙련의 균형을 강하게 요구하는 것으로 나타났다. 따라서 향후 직업훈련과정에서는 장비 선택, 수리, 조작과 관련된 인지적 스킬 및 신체적 균형을 중심으로 한 훈련이 요구된다고 할 수 있다.	Gyuhwan Kim, Yeongbin Choi, Junemo Koo
8	2025	Equalising the effects of automation? The role of task overlap for job finding ↗ This paper directly addresses the distributional effects of automation and AI exposure, focusing on how task overlap influences labor market outcomes for displaced workers. It aligns closely with the project's themes on inequality, occupation-specific impacts, and the distinction between different types of technological exposure. This paper investigates whether task overlap can equalise the distributional effects of automation for unemployed job seekers displaced from routine jobs. Using a language model, we establish a novel job-to-job task similarity measure. Exploiting the resulting job network to define job markets flexibly, we find that only the most similar jobs affect job finding. Since automation-exposed jobs overlap with other highly exposed jobs, task-based reallocation provides little relief for affected job seekers. We show that this is not true for more recent software exposure, for which task overlap lowers the inequality in job finding. • Similar jobs offer little relief from reduced job finding due...	Diego Dabed, Sabrina Genz, Emilie Rademakers
8	2025	Artificial Intelligence in the Office and the Factory: Evidence from Administrative Software Registry Data ↗ [Title only] This paper directly addresses firm-level AI adoption and the resulting reorganization of work by leveraging unique administrative data to distinguish between office and factory settings. Its focus on measuring exposure and observing real-world implementation aligns closely with the core themes of AI exposure measurement and task-based frameworks. No abstract available.	Gustavo de Souza
8	2025	AI-Induced Labor Market Shifts in the U.S.: Occupational Exposure and Regional Disparities ↗ This paper directly addresses core project themes by developing novel metrics for AI exposure (replacement vs. assistive) and empirically estimating their causal effects on wages and employment using a Difference-in-Differences approach. It provides crucial regional and occupational context for understanding distributional effects and the augmentation versus substitution dynamics of AI in the labor market. Abstract This study explores how Artificial Intelligence (AI), particularly the rise of Large Language Models (LLMs), is unevenly transforming the U.S. labor market. Focusing on wage growth and employment patterns across various occupations and regions, we introduce two new indices—the Replacement Exposure Index (REI) and the Assistive Exposure Index (AEI)—to measure the susceptibility of jobs to automation or enhancement by AI. Using panel data from the U.S. Bureau of Labor Statistics and the O*NET Resource Center, we conduct a Difference-in-Differences (DID) analysis with the emergence of LLMs in 2022 as a key turning point. Our findings indicate that occupations exposed to AI experience...	Jiamei Wang, Dongyan Wang
8	2025	Generative AI and Jobs: An Analysis of Potential Effects on Global Employment ↗ This paper directly addresses the project's core questions by measuring AI exposure at the task level and analyzing its potential distributional effects on employment across income groups and demographics. It provides relevant evidence on whether AI acts as an augmentor or substitutor, highlighting significant variations in exposure by gender and country income level. Badanie przedstawia globalną analizę potencjalnego wpływu generatywnej sztucznej inteligencji na zatrudnienie. Korzystając z modelu GPT-4, szacujemy wyniki na poziomie zadań i oceniamy ich potencjalne oddziaływanie na zatrudnienie w skali globalnej oraz w krajowych grupach dochodowych. Stwierdzamy, że praca biurowa jest jedyną spośród szerokich kategorii zawodowych w wysokim stopniu wystawionych na oddziaływanie technologii, podczas gdy inne grupy zawodowe, takie jak menedżerowie czy specjaliści, wykazują znacznie niższe poziomy ekspozycji. W związku z tym główny wpływ generatywnej sztucznej inteligencji będzie prawdopodobnie polegał na usprawnieniu pracy, a nie na pełnej automatyzacji...	Paweł Gmyrek, Janine Berg, David Bescond
8	2025	Occupational exposures, complementarity and the potential consequences of A.I. for the labour market: some evidence from Ireland ↗ This paper directly addresses the core theme of measuring AI exposure across occupations and analyzing its distributional effects by skill, age, and gender in the Irish labor market. It provides empirical evidence on AI complementarity and potential labor market disruption, aligning closely with the project's focus on task-based frameworks and inequality. The adoption of AI technology by industry could significantly disrupt our current understanding of “typical” economic activity. As AI comes to pervade more sectors and occupations over time, it is likely that this technology will give rise to challenges and risks but also opportunities and benefits. There is, however, a significant degree of uncertainty regarding how future waves of technological change will impact the economy, including the labour market. Recent research has found that 40% of employment globally is exposed to AI and that this rises to 60% of employment in advanced economies. We analyse exposure and complementarity in tandem in order to better understand the potential...	Harry Williamson, Dermot Coates, Kevin Daly, et al.
8	2026	The Impact of Generative Artificial Intelligence on Market Equilibrium: Evidence from a Natural Experiment ↗ This paper closely relates to the project by empirically examining how generative AI affects labor market outcomes, specifically wages and employment volume, through a rigorous natural experiment design. It provides direct evidence on whether AI acts as a substitute or complement for creative labor and how task reorganization impacts market equilibrium, aligning with the core themes of causal effects on wages and distributional outcomes. Generative artificial intelligence (AI) exhibits the capability to generate creative content akin to human output with greater efficiency and reduced costs. This groundbreaking capability, however, has ignited a debate regarding its potential to displace human creators. In light of these discussions, this paper empirically investigates the impact of generative AI on market equilibrium, in the context of China's leading art outsourcing platform. We overcome the challenge of causal inference by identifying an unanticipated and sudden leak of an advanced image-generative AI as a natural experiment. This leak precipitated a notable reduction in the production costs of anime-style images...	Kaichen Zhang, Zixuan Yuan, Hui Xiong
8	2026	In-demand skills: a shield against automation—evidence from online job vacancies ↗ This paper directly addresses the core themes of AI exposure measurement and skill complementarity by constructing a novel AI exposure metric and analyzing its correlation with in-demand skills. It provides valuable empirical evidence on how specific skill bundles act as shields against automation, contributing to the understanding of which workers are winners and losers in the AI transition. This paper investigates how in-demand skills, advertised wages, and occupational exposure to automation co-evolve in Slovakia’s online labor market. Drawing on data covering nearly the full universe of online job vacancies posted in Slovakia in 2022, the analysis extracts skills from unstructured text and maps them into fifteen conceptual categories spanning cognitive, socio-emotional, and manual domains. These categories account for a sizeable share of wage variation, with several linked to notable premia or penalties. Automation risk is gauged through a novel Europe-specific measure of exposure to AI and machine learning, software, and robotics, constructed by matching patent text to...	Tomáš Oleš
8	2026	The risks and bottlenecks to automation in employment in Argentina. New impacts on the occupational structure in selected economic sectors ↗ The paper directly addresses the project's core theme of measuring AI exposure and risk across occupations by constructing an automation risk index based on job skills and tasks in an emerging economy context. It provides valuable empirical evidence on skill complementarity versus substitution, highlighting how AI is currently used to augment rather than replace complex, skilled tasks, which informs the distributional effects of AI on labor markets. = 426) was conducted using the Respondent Driven Sampling (RDS) technique. The analysis covered sectors with different levels of technological integration-high (e.g., software), medium (e.g., food), and low (e.g., textiles). An automation risk index was constructed based on job skills and tasks. The results indicate that professionals, scientists, managers, and technicians exhibit a lower risk of automation, while elementary and industrial occupations face a higher risk. Social and creative intelligence were identified as 'bottlenecks' in the face of automation, an aspect that we have emphasised in this analysis. The software and pharmaceutical sectors are more protected, unlike the textile...	Eduardo Chávez Molina, José Rodríguez de la Fuente
8	2026	AI‐powered skill classification: mapping technology intensity in the German labour market ↗ This paper directly addresses the project's core theme of measuring AI exposure by developing a novel metric for occupational technology intensity that distinguishes AI from other technologies. It provides crucial empirical context by linking these measures to administrative data to reveal how AI adoption correlates with employment trends across different skill levels. Abstract The rapid evolution of technology is reshaping labour markets by altering skill demands and job profiles. This paper introduces a novel skill‐based measure of occupational technology intensity – the occupational technology skill share (OTSS) – that distinguishes between manual, digital and frontier technologies, including artificial intelligence (AI). Using natural language processing, generative AI and supervised machine learning, we develop an AI‐powered skill classification that enriches occupation‐linked skill labels with standardised GenAI‐generated descriptions and structured indicators of technological content, enabling transparent classification by technology intensity. We...	Sabrina Genz, Terry Gregory, Florian Lehmer
8	2026	Toward Sustainable Workforce Development: How AI Reshapes Skill Demand Structure—Evidence from 67 Million Job Postings in China ↗ This paper directly addresses the project's core themes by analyzing how AI reshapes skill demand structures through a task-based decomposition into displacement and augmentation effects. It provides valuable empirical evidence on which skill categories are complemented or substituted by AI, informing the distributional effects and firm-level reorganization questions central to the researcher's project. How artificial intelligence (AI) reshapes the internal structure of firm-level skill demand remains largely uncharted. Using approximately 67 million online job postings from two major Chinese recruitment platforms (2019–2024), we construct firm-by-year potential AI exposure via semantic matching between AI patent texts and detailed occupation task descriptions, decompose exposure into displacement and augmentation components based on task routineness, and measure four skill-category demand shares and their within-category importance from job-description text, with identification from within-firm variation under firm and city-by-year fixed effects. Displacement and augmentation exposure...	Ling Zhang, Chenglei Zhang
8	2026	Artificial Intelligence and Work Intensity: Evidence From Chinese Listed Firms ↗ The paper directly investigates the causal effects of AI on labor market outcomes by analyzing its impact on work intensity, a key dimension of worker productivity and conditions. It employs rigorous methods to identify substitution and complementarity effects while exploring heterogeneity across firm types and regions, aligning closely with the project's themes on AI's labor market reshaping and distributional effects. ABSTRACT Artificial intelligence (AI) is reshaping labor markets at an unprecedented pace. Existing studies on the impact of AI on labor markets primarily focus on extensive margin effects, while its influence on intensive margin labor remains underexplored. This study empirically examines the effect of AI exposure on work intensity in Chinese publicly listed firms by integrating satellite nighttime lights, firm employee occupational structures, and occupation‐level AI exposure data. The results show that AI exposure significantly increases firm work intensity. This finding remains robust after using an instrumental variable approach constructed from AI patent text analysis. Substitution...	Lilong He, Xiangyang Chen, Juan Liu
8	2025	Adoption of artificial intelligence in Spanish firms: an initial analysis based on the Banco de España Business Activity Survey ↗ This paper directly addresses the core theme of firm AI adoption, providing empirical evidence on which sectors and firm characteristics drive uptake. It also offers relevant insights into the specific tasks AI is applied to and how firms perceive its impact on employment, which are central to the project's research questions. Rationale Artificial intelligence (AI) has the potential to revolutionise economies and labour markets. Using the Banco de España Business Activity Survey (EBAE), this article analyses the adoption of AI by Spanish firms. Takeaways •Almost 20% of the Spanish firms surveyed are using AI systems. This is less than their German counterparts, but higher than the adoption rate in Italy, according to similar surveys conducted in those countries. However, most firms are still just experimenting with AI. •The AI adoption rate is higher in technology services and in large, productive and young firms. The main obstacles are the lack of skilled labour, high implementation costs and data availability...	Alejandro Fernández Cerezo, I. Vallinas Hidalgo, Mario Izquierdo
8	2025	Artificial Intelligence Technologies and Employee Pay in the United Kingdom: Evidence From Matched Employer–Employee Data ↗ This paper directly addresses the project's core theme of distributional effects by empirically analyzing how AI adoption impacts employee wages across different skill and occupation groups in the UK. It provides relevant evidence on whether AI acts as a complement or substitute for lower-skilled labor, contributing to the understanding of wage inequality and the identification of winners and losers in the labor market. ABSTRACT This paper examines the impact of artificial intelligence (AI)‐enabled technologies on employee pay in the United Kingdom. We use matched nationally representative data from the Employers’ Digital Practices at Work Survey and an original survey of 6000 UK workers and apply machine learning techniques to uncover relationships between AI technology and employee pay across qualification and occupation skill groups. We find that lower skilled workers were the primary beneficiaries of AI, but this effect was contingent on the extent of worker interaction with AI. Further analysis shows that employee involvement in pay determination facilitates a more equitable distribution of AI‐related...	Felix Schulz, Danat Valizade, Mark Stuart, et al.
8	2025	Beyond substitution: AI’s reshaping of low-skilled labour demand in China - firm-level evidence from job posting data ↗ [Title only] This paper directly addresses the core themes of AI's impact on labor markets by focusing on low-skilled workers and using firm-level job posting data to analyze labor demand changes in China. It likely provides valuable empirical evidence on whether AI substitutes for or reshapes demand for low-skill labor, contributing to the understanding of distributional effects and firm-level responses. No abstract available.	Junming Zhuang, Chen Yang, Yeqing Ma
8	2026	Training for Technology: Adoption and Productive Use of Generative AI in Legal Analysis ↗ This paper directly addresses the project's core theme of generative AI productivity experiments by isolating the causal impact of user training on LLM adoption and performance. It provides valuable empirical evidence on how complementary investments in human capital influence the productive use of AI in knowledge-intensive occupations, a key mechanism in understanding AI's effect on worker productivity and skill complementarity. Can targeted user training unlock the productive potential of generative artificial intelligence (GenAI) in professional settings? We investigate this question using a randomized study involving 164 law students completing an issue-spotting examination. Participants were assigned to one of three conditions: no GenAI access, optional access to a large language model (LLM), or optional access accompanied by an approximately ten-minute training intervention.<br><br>Training significantly increased LLM adoption—the usage rate rose from 26% to 41%—and improved examination performance. Students with trained access scored 0.27 grade points higher than those with untrained access (p = 0.027)...	Benjamin Minhao Chen, Hong Bao
8	2026	Who Adopts Matters: Generative AI and Short-Run Wage Compression ↗ This paper directly addresses the project's interest in distributional effects and the winners/losers of AI adoption by analyzing how asymmetric adoption compresses wage premiums. It provides specific empirical evidence on skill-based inequality and challenges standard aggregation methods used to assess AI's labor market impact. Abstract Generative AI is diffusing unevenly across the workforce. Policymakers rely on aggregate productivity estimates to guide resource allocation, yet standard linear aggregation implicitly assumes perfect substitutability between worker types. We embed the micro-data of Bick et al. (2026) in a constant elasticity of substitution (CES) framework ($\:\sigma\:=1.4$) to examine distributional consequences that the linear approach cannot capture. Three findings emerge. First, the aggregate productivity estimate is robust: the CES-linear gap is below 0.01 percentage points. Second, asymmetric adoption compresses the skill-service price premium by 0.65%, scaling to approximately 2% under...	Zhihao Sun, Chun-Chieh Hu
8	2026	Role of Complementary Capabilities in Artificial Intelligence Adoption and Productivity: Firm-Level Evidence from Canada ↗ This paper directly addresses the project's theme of how firms reorganize work and the conditions under which AI augments labor through complementary capabilities like IT infrastructure and R&D. It provides firm-level empirical evidence on AI adoption patterns and productivity effects, which is crucial for understanding the causal mechanisms behind AI's impact on labor markets. L’étude examine les facteurs associés à l'adoption de l'intelligence artificielle (IA) par les entreprises exploitées au Canada, ainsi que les effets connexes sur la productivité, en s'appuyant sur une nouvelle base de données d'entreprises reliant l'Enquête sur la technologie numérique et l'utilisation d'Internet et des données organisationnelles. À l'aide d'une gamme de techniques économétriques, notamment l'estimation par écart dans les différences, nous évaluons la corrélation entre l'adoption de l'IA et la productivité du travail des entreprises. Nos résultats montrent que l'adoption de l'IA est étroitement liée à des capacités complémentaires, comme des investissements en recherche et...	Jiang Li, Huju Liu
8	2026	How artificial intelligence penetration affects workers’ job quality? Empirical evidence from China ↗ This paper directly addresses the project's core themes by empirically measuring AI penetration's causal effects on worker outcomes, specifically focusing on job quality and distributional impacts across skill and gender groups. It aligns closely with the research goals by examining task-based mechanisms and distinguishing between augmentation effects, providing valuable evidence on the winners and losers in the AI-driven labor market. Artificial intelligence (AI) has evolved to play a transformative role in reshaping labor market patterns, altering how individuals work and the tasks they perform. The study investigates how AI penetration influences workers’ job quality and explores the underlying mechanism of this effect. To more accurately capture the extent of AI application, we establish a measure of AI penetration by applying natural language processing (NLP) techniques in combination with the Bartik instrumental variable approach. Job quality is measured using a multidimensional index. Employing nationally representative data from the China Family Panel Studies, we find that AI penetration significantly improves job...	Yafeng Nie, Kaiyun Xue
8	2026	The impact of artificial intelligence on labor income share: From the perspective of technological progress bias ↗ This paper directly addresses the aggregate labor market effects of AI by analyzing its impact on the labor income share, a key component of wage inequality and distributional outcomes. It employs a robust empirical strategy to examine how AI induces capital-biased technological change, offering critical insights into the macroeconomic consequences of AI adoption on worker compensation. Artificial intelligence (AI), a general-purpose technology with the characteristics of a new type of infrastructure, can augment both capital and labor. Its impact on the labor income share (LIS) remains debated. From the perspective of biased technological progress, this paper investigates how AI affects the LIS through both theoretical and empirical analyses. First, using a factor-augmenting CES production function, we incorporate AI and biased technological progress into a unified framework for the determination of the LIS and clarify the sources of variation in labor's share. Second, we use machine learning to build an AI dictionary, perform patent text analysis, and construct a...	Lei Chen, Fang He, Yongxue He, et al.
8	2026	AI adoption among German firms ↗ This paper closely relates to the project by providing empirical evidence on firm-level AI adoption, a core theme, while addressing specific drivers like managerial traits and sectoral diffusion patterns. Although it focuses on potential rather than realized productivity impacts and lacks direct labor market outcome data, it offers valuable context on the adoption mechanisms that precede the labor market effects of interest. This paper examines the adoption of Artificial Intelligence (AI) among German firms, leveraging firm-level data from the ifo Business Survey. We analyze the diffusion of AI across sectors and firm sizes, showing a significant increase in AI usage from 2023 to 2024, particularly in manufacturing and services. The survey data allows us to explore not only sectoral patterns of adoption but also the drivers and barriers that firms face, including firm-specific characteristics and industry dynamics. Additionally, we investigate the role of managerial traits, such as risk tolerance and patience, in shaping AI adoption decisions. Finally, we assess the potential productivity impacts of AI at the...	Thomas Licht, Klaus Wohlrabe
8	2026	The Jevons trap: when artificial intelligence in healthcare creates endless work ↗ [Title only] This paper directly addresses the core themes of task reorganization and whether AI augments or substitutes for labor, specifically examining the paradoxical increase in workload (endless work) despite efficiency gains. It provides critical insight into the distributional effects and firm-level responses to AI adoption within a specific high-skill sector. No abstract available.	Ravi Shankar
8	2024	Mapping Workforce Mobility: Skills, Data and Measurement ↗ This paper is closely related to the project as it develops a novel, data-driven method for measuring workforce skills and AI exposure through job posting embeddings. It provides critical empirical insights into how technological change drives worker mobility and wage outcomes, directly addressing the project's themes of measurement, distributional effects, and labor market adaptation. The interplay of skills, technological change, and worker mobility is central to understanding evolving labor markets. This paper enhances our understanding of worker mobility by introducing a novel, data-driven approach to mapping occupational skills, constructing occupation-specific skill embeddings from job posting-derived prompts. This improved ability to account for worker mobility, compared to traditional measures, allows us to analyze the distinct impacts of technological change: both technology enhancing tasks by requiring new skills, and technology eliminating tasks. Combining this enhanced skill measure with technology exposure metrics and worker transition histories (2002-2018)...	Nikhil George, Christophe Combemale, Ramayya Krishnan, et al.
8	2025	Technological change and insuring job loss ↗ This paper directly addresses the project's core theme of how technological change affects workers by quantifying its role in earnings declines following job loss. It provides critical empirical evidence on skill mismatch and occupational mobility, which are key mechanisms in understanding the distributional effects and potential substitution impacts of AI in labor markets. We examine the role of technological change in explaining the large and persistent decline in earnings following job loss. Using detailed skill requirements from the near universe of online vacancies, we estimate technological change by occupation and find that technological change accounts for 45 percent of the decline in earnings after job loss. Technological change lowers earnings after job loss by requiring workers to have new skills to perform newly created jobs in their prior occupation. When workers lack the required skills, they move to occupations where their skills are still employable but are paid a lower wage. (JEL J24, J31, J63, O33)	J. Carter Braxton, Bledi Taska
8	2025	Measuring the Uncertain Task Content of Work: Workforce Design for AI-Driven Human-Robot Collaboration in Intralogistics ↗ This paper directly addresses the core theme of measuring AI exposure by developing a novel stochastic task-level framework to estimate full-time equivalents based on task frequency and AI exposure scores. It provides a rigorous methodological contribution to the project's interest in how AI exposure varies across occupations and tasks, specifically within the context of workforce planning and human-robot collaboration. This paper addresses the challenge of strategic workforce planning for AI-driven human-robot collaboration (AI-HRC) in intralogistics. We ask two questions: how can task-level full-time equivalent (FTE) estimates be constructed from existing labor statistics, and how can these estimates, combined with AI exposure metrics, inform strategic AI-HRC design and workforce planning? Drawing on U.S. Bureau of Labor Statistics employment data, O*NET occupational profiles, and task-level AI exposure scores, we develop a stochastic task-time framework that decomposes occupations into tasks and models task frequencies as probability vectors on the simplex. A covariance-completion procedure reconstructs...	Pierre Bouquet, Nicolo Bagnoli, Yossi Sheffi
8	2026	Industries most exposed to AI are not only seeing productivity gains but jobs and wage growth too ↗ [Title only] The title directly addresses the core questions regarding the causal effects of AI on employment and wages, as well as the aggregate labor market implications of AI adoption. It specifically targets the debate on whether AI acts as a substitute or complement, focusing on the industries most exposed to these technologies. No abstract available.	Christos Makridis
8	2026	Artificial intelligence and firms’ labor income share: measurement and analysis based on LLM ↗ [Title only] This title directly addresses the project's interest in how AI reshapes labor markets by focusing on the labor income share, a key aggregate distributional metric. The use of LLMs for measurement aligns with the project's theme of developing new methods to assess AI exposure and its economic impacts. No abstract available.	Tianjian Huang, Liang Wang, Peiqi Yang
8	2024	Exploring the Dual Impact of AI on Employment and Wages in Chinese Manufacturing ↗ This paper directly addresses the core question regarding the causal effects of AI on employment and wages, specifically focusing on the manufacturing sector. It provides empirical evidence on the distributional impacts by analyzing heterogeneity across firm size and region, which aligns with the project's interest in identifying winners and losers. Purpose- This study investigates AI's impact on employment and wage dynamics within the manufacturing sector.Design/Methodology- Utilizing data from 3,522 manufacturing firms between 2007 and 2021, we analyze the effects of AI adoption on labor markets. Findings- AI adoption correlates with reduced employment numbers yet enhances wage rates, with some employees seeing wage increases as high as 83.86%. Heterogeneity analysis reveals variability in these impacts, dependent on contextual factors. The deployment of artificial intelligence in manufacturing sectors leads to an upgraded wage structure, emphasizing the importance of advancing individual professional skills to capitalize on these...	Jing Wang
8	2025	Return of the Solow-paradox in AI?: AI-adoption and firm productivity ↗ This paper directly addresses the core theme of aggregate labor market effects by empirically estimating the causal impact of AI adoption on firm productivity. It provides valuable context for understanding distributional winners and losers by highlighting that productivity gains are concentrated in large firms and subject to significant implementation delays. Abstract AI-related technologies have become ubiquitous in many contexts within the business world. However, to date, empirical accounts of the productivity effects of AI adoption are scarce. Using data from Finnish job advertisements posted between 2013 and 2019, we identify job advertisements referring to AI-related skills. Matching this data to productivity data from ORBIS, we then estimate the productivity effects of AI-related activities in our sample. Our results indicate that AI-adoption increases productivity—with three important qualifications. Firstly, effects are only observable for large firms with more than 499 employees. Secondly, there is evidence that early adopters did not...	Arash Hajikhani, Angela Jäger, Torben Schubert, et al.
8	2025	Impact of Artificial Intelligence on the Future of Education, Skill Development and Job Market: Blessing or a Curse ↗ This paper directly addresses the core theme of AI exposure measurement by constructing an occupation-wise AI Susceptible Index using OECD and World Bank data. It provides empirical evidence on how AI impacts vary by skill level and occupation, contributing to the understanding of which workers are most likely to be substituted or affected by generative AI. Of late, there has been a revolutionary shift in technology, digitalization and machine learning which has opened up a new frontier of ‘generative artificial intelligence’ (AI), across the globe. AI basically refers to a spectrum of interrelated machine learning technologies applied to solve problems inspired by human intelligence. Future of Jobs Report (May 2023) has predicted a churn in the structure and composition of the labour market. One may decompose the effect of this AI into three broad dimensions: a displacement effect, a productivity effect and a reinstatement effect. The present chapter attempts to investigate the possible divergent impact of AI across countries (developed...	Rajib Bhattacharyya, Kaushiki Banerjee
8	2025	An Assessment of Occupational Exposure to Artificial Intelligence in Italy ↗ [Title only] This paper directly addresses the core theme of measuring AI exposure across occupations, providing critical granular evidence from Italy that complements broader global studies. Its task-based assessment of occupational risk is highly relevant to understanding how AI reshapes labor markets and identifying specific groups of workers who may be affected. No abstract available.	Antonio Dalla Zuanna, Davide Dottori, Elena Gentili, et al.
8	2025	IT investment and labor displacement: Evidence from Korea ↗ This paper directly addresses the core question of whether AI and IT investments substitute for labor, providing empirical evidence on the causal relationship between technology adoption and workforce displacement. It further explores the distributional and aggregate effects by analyzing changes in wages, productivity, and the labor share, which aligns closely with the project's focus on firm reorganization and macroeconomic impacts of AI. There is widespread concern that many companies heavily invest in IT, such as artificial intelligence (AI) and cloud technologies, in order to replace human labor. Using novel Korean data, we verify whether such an effect is present in the cross-sectional variation among firms’ IT investment and workforce size. At the firm level, the economy-wide effect of doubling IT investment per worker is associated with a reduction of over 60 workers in total workforce size. Meanwhile, IT investment seems to be enhancing productivity, as reflected in its positive association with both the total salary bill and the average wage per capita. The overall effect of IT investment appears to be a decrease in...	Joon‐Bae Lee
8	2025	Artificial Intelligence Upskills Human-Centric and Fundamental Computer Science Knowledge ↗ This paper directly addresses the project's core themes of AI exposure measurement and AI skill complementarity by analyzing shifts in labor demand across occupations. It provides valuable empirical evidence on how AI adoption is reshaping required skills, specifically highlighting the rising importance of human-centric and fundamental computer science knowledge. What skills should workers acquire to remain valuable in a labor market being transformed by artificial intelligence (AI)? This is a crucial question for workers to adapt to the workforce disruptions driven by AI adoption. We address this question by identifying the job skills that are in increasing demand as AI applications expand across occupations. Specifically, we measure occupational AI exposure by considering the textual similarity between occupational descriptions and AI patents. Then, based on 347 million job postings in the US, we consider 442 distinct skills and examine how the demand for skills changes with occupational AI exposure from 2010 to 2023. We find that occupations with...	Pedro Masi, Yuanyang Liu, Chuanren Liu
8	2026	Human Capital Investment in the Age of Artificial Intelligence ↗ This paper closely addresses the project's theme of entry-level workers and the traditional career ladder by analyzing how AI exposure influences human capital investment and degree completion patterns. It provides relevant empirical context on how young workers may be adjusting their educational trajectories in response to generative AI risks. <div> This paper documents degree completion patterns in programs associated with occupations highly exposed to generative AI and examines the educational background composition of workers in these occupations. Controlling for school-level shocks, event-study estimates show that by 2024 the number of bachelor’s degrees conferred in programs associated with high occupational AI exposure increases by 5% relative to programs associated with low occupational AI exposure. Programs linked to occupations where AI is more likely to automate tasks also experience a significant increase in bachelor’s degrees conferred. This increase is most pronounced at public and R1 institutions. So far...	Zhengyi Yu
8	2026	Causal Identification of Artificial Intelligence Effects on Enterprise Labor Structure via a Partially Linear Double Machine Learning Estimator: Evidence from High-Dimensional Panel Data ↗ This paper directly addresses the project's core theme of AI skill complementarity versus substitution by providing causal evidence that AI adoption increases the share of high-skilled labor in Chinese firms. Its application of advanced econometric methods to firm-level panel data offers valuable empirical insights into how AI reshapes enterprise labor structures, aligning closely with the research's focus on distributional effects and firm reorganization. This study develops a semiparametric causal inference framework to quantify the effect of Artificial Intelligence (AI) adoption on enterprise labor structure under high-dimensional confounding. We employ the Double Machine Learning (DML) estimator proposed, which combines Neyman orthogonality and cross-fitting to achieve reliable causal identification in settings where conventional regression methods are prone to bias from high-dimensional controls and nonlinear confounding. Nuisance functions are estimated using Lasso and Random Forests, enabling flexible modeling of complex relationships between control variables and outcomes. Using an unbalanced panel of Chinese A-share listed companies...	Huali Liu, Wenjie Li, Yankai Lin, et al.
8	2026	ДВУСТОРОННЯЯ ЛОВУШКА НЕДОИНВЕСТИРОВАНИЯ В ИИ-КОМПЕТЕНЦИИ: ФОРМАЛЬНАЯ МОДЕЛЬ ↗ This paper directly addresses the project's themes of AI skill complementarity, worker retraining, and the coordination challenges firms and employees face in adapting to AI depreciation. By modeling the strategic interdependence in investment in AI competencies, it provides a theoretical framework relevant to understanding who is affected by AI exposure and how labor market dynamics shift. В работе строится формальная двухагентная модель инвестиций в обучение маркетологов в условиях ускоренной ИИ-депрециации человеческого капитала. В модели работодатель и работник одновременно выбирают доли инвестиций в обучение компетенциям с заданной скоростью депрециации δ и горизонтом окупаемости T. Выводится аналитическое условие прибыльности инвестиции δ·T ≤ ln 2. Внутри этой области индивидуально рациональное равновесие совпадает с социальным оптимумом, а в промежуточной зоне ln 2 < δ·T ≤ ln 2 + ln(1 + γ/α) стратегическая взаимодополняемость инвестиций порождает координационную ловушку: совместное вложение выгодно обоим агентам, но ни один не начинает первым. Показано, что для трёх...	В.Г. Абрамян, Е.В. Мищенко
8	2026	Generative <scp>AI</scp> : Catalyst for Growth or Harbinger of Premature De‐Professionalization? ↗ This paper directly addresses the project's core themes by modeling the macroeconomic and distributional effects of generative AI on labor markets, specifically focusing on skill complementarity and the potential for 'premature de-professionalization.' It provides a theoretical framework for understanding how AI reshapes labor demand across high-skill tradable services versus low-skill sectors, offering relevant insights into inequality and structural transformation. ABSTRACT This paper develops a multi‐sector growth model to analyse the general equilibrium effects of generative artificial intelligence (GenAI) on growth, structural transformation and international trade. It introduces a crucial distinction between high‐skill, high‐income, highly digitalized, tradable services and low‐skill, less digitalized, less‐tradable services, allowing GenAI to differentially affect sectors through both supply‐side productivity gains and demand‐side shifts in consumer preferences. The model generates testable implications consistent with emerging empirical evidence and delivers several novel predictions. Unless GenAI adoption is broad‐based and accompanied by the...	Yan Liu
8	2022	The effect of early automation on the wage distribution with endogenous occupational choices ↗ This paper provides crucial historical context for understanding how automation affects wage distribution and occupational choices, which is central to the project's inquiry into AI's distributional effects. It employs a task-based framework with endogenous worker choices, directly informing the theoretical mechanisms regarding whether technology substitutes for or complements labor across different skill levels. Abstract While the literature demonstrated that automation reduces employment in routine jobs (job polarization), its impact on wages is still unclear and the debate open. By applying Counterfactual Quantile Regressions to historical data, this paper analyzes the channels through which automation affected wage inequality in the U.S. labor market during the 1990s. Contrary to conventional wisdom, we find that the observed decline in wage inequality among low earners was not due to lower prices paid for technology-substitute occupational tasks, but instead due to more homogeneous wages of workers performing these tasks. This evidence is consistent with a model of directed (routine-biased)...	Manuel Hidalgo-Pérez, Benedetto Molinari
8	2026	Will Human Creatives Survive Generative AI Shocks? ↗ This paper directly addresses the project's core question of whether AI tools augment or substitute for labor, specifically within the creative sector. It provides a theoretical framework for task complementarity and human-AI coexistence, offering testable predictions on how AI exposure affects employment dynamics in creative occupations. Concerns are mounting about the impact of generative AI on human jobs: With ongoing improvements, would generative AI models eventually replace all human creatives one day? Inspired by the familiar insight on the impossibility of an informationally efficient financial market [e.g., Grossman and Stiglitz (1980)], we argue that human content creation remains essential despite the rise of generative AI. We formalize our insight in a series of models with increasing complexity: First, a stylized static model shows robust human-AI coexistence: If, hypothetically, generative AI models produce all content we need at low variable costs, then there would be no incentive for humans to spend costly...	Bo Hu, Jiasun Li
8	2026	Can ChatGPT Kill User-Generated Q&A Platforms? ↗ This paper directly addresses the project's theme of task reorganization and the substitution versus complementarity of AI by analyzing how LLMs displace routine knowledge tasks on Q&A platforms. It provides empirical evidence on the specific occupational impacts of generative AI on user-generated content, highlighting shifts in task complexity and the resulting dynamics between human experts and AI systems. Large language models (LLMs), such as ChatGPT, exhibit substantial functional overlap with user-generated knowledge ecosystems while also relying on them as critical inputs for future learning. This dual role creates a fundamental tension that calls for a clearer understanding of how LLMs reshape these ecosystems. Adopting a niche theory perspective, we examine how functional overlap and knowledge structure determine the boundary between substitution and coexistence. Using Stack Overflow, we show that LLM introduction reduces question volume by about 14% on average (and up to 27.9% over time), with stronger declines in mid- to low-quality content, in topics with richer and more structured...	Junzhi Xue, Lizheng Wang, Jinyang Zheng, et al.
8	2025	Labor Market Adjustment to Generative AI: The Good, the Bad, and the Ugly ↗ [Title only] The title explicitly addresses labor market adjustments driven by generative AI, directly aligning with the project's core themes of causal effects on employment and wages. The inclusive phrasing 'The Good, the Bad, and the Ugly' suggests a comprehensive analysis of distributional winners and losers, which is central to the research scope. No abstract available.	Nima Rafizadeh
8	2025	The Twin Disruption of Generative AI: Unbundling Capability and Responsibility in Creative Work ↗ [Title only] This paper likely investigates how generative AI transforms creative labor markets by separating technical capability from accountability, directly addressing the project's focus on task reorganization and labor market effects. It promises insights into who wins or loses in the AI era by examining the shifting responsibilities and skill demands within creative occupations. No abstract available.	Emmanuelle Walkowiak, Jason Potts
8	2026	AI and the democratization of knowledge work ↗ [Title only] This title directly addresses the project's interest in how AI reshapes labor markets, specifically by potentially lowering barriers to entry for knowledge work tasks traditionally held by higher-skilled workers. It likely explores the distributional effects and task reorganization aspects of generative AI, which are core themes of the research project. No abstract available.	Madeleine I. G. Daepp, Kiran Tomlinson, Scott Counts, et al.
8	2026	Data and Code for: Not a Typical Firm: Capital-Labor Substitution and Firms' Labor Shares ↗ This paper directly addresses the project's core theme of how firms reorganize work and the distributional effects of automation on labor shares by modeling capital-labor substitution. It provides relevant empirical context on heterogeneity in firm-level automation adoption and its macroeconomic implications for wages and inequality. Replication package for "Not a Typical Firm: Capital-Labor Substitution and Firms' Labor Shares". <br>Abstract: "The US labor share has declined, especially in manufacturing and retail. Yet, the labor share of a typical firm in these sectors has risen. We introduce a model where firms incur fixed costs to automate tasks. A decline in the price of capital goods used for automation reproduces the observed patterns: large firms automate tasks, reducing the aggregate labor share, while the median firm continues to operate a labor-intensive technology. When calibrating the automation fixed costs to match the observed adoption heterogeneity, the model generates the aggregate and firm-level facts...	Joachim Hubmer, Pascual Restrepo
8	2026	<p>Resolving the Automation Paradox: Falling Labor Share, Rising Wages</p> ↗ [Title only] This title suggests a macroeconomic analysis of labor market outcomes related to automation, directly addressing core themes of falling labor shares and wage effects. It likely provides aggregate evidence on whether automation acts as a substitute or complement to labor, fitting the researcher's interest in economywide employment and wage effects. No abstract available.	David Autor, BN Kausik
8	2025	Job Transformation, Specialization, and the Labor Market Effects of AI ↗ [Title only] The title explicitly addresses labor market effects of AI, job transformation, and specialization, which directly aligns with the project's core themes of causal effects, task-based frameworks, and firm reorganization. While the mention of specialization suggests a focus on distributional winners and losers, the lack of specific keywords like 'generative' or 'productivity experiments' introduces slight uncertainty regarding its exact empirical scope. No abstract available.	Lukas B. Freund, Lukas Friedrich Mann
8	2026	<scp>AI</scp> Skills Wanted: How <scp>AI</scp> Technologies Create Demand for Skilled Workers ↗ This paper directly addresses the project's core theme of how AI reshapes labor markets by creating demand for specific skilled workers, aligning with the investigation into skill complementarity and distributional effects. It provides relevant empirical evidence using job-posting data to document the rise in AI skill demand, offering valuable context for understanding the changing composition of occupations. History has shown that major technological innovations reshaped the skill composition of occupations. Artificial intelligence (AI) is driving a similar transformation today. This chapter begins by showing, through examples of the automobile and computer revolutions, how general-purpose technologies have historically redefined existing jobs and created entirely new ones. Using international job-posting data, the chapter documents the rapid and broad-based rise in demand for AI skills—both technical and AI literacy. The introduction of computers in the mid-20th century offers another vivid example of how technological change, driven by a general-purpose technology, can generate entirely new...	Fabian Stephany, Ole Teutloff
8	2024	Generative AI & TeamWork: An experimental approach. ↗ [Title only] This paper directly addresses the core theme of generative AI productivity experiments by investigating how AI tools impact team dynamics and performance. It provides insight into how firms may reorganize work and whether AI acts as a complement or substitute for labor in collaborative settings. No abstract available.	Fabrizio Dell’Acqua, Charles Ayoubi, Karim R. Lakhani
8	2025	Analysis of Artificial Intelligence Exposure Across Industries in South Korea and the United States ↗ [Title only] This paper directly addresses the core theme of AI exposure measurement by comparing its variation across industries in two major economies. The cross-country comparative approach offers valuable insights into how institutional or structural differences might influence AI adoption and occupational exposure, which is central to understanding distributional effects. No abstract available.	Yaein Baek, Jiyun Lee
8	2026	The Ironies of Automation: A Growth Theory Perspective ↗ This paper directly addresses the project's core themes of AI labor market effects, skill complementarity versus substitution, and aggregate economic impacts through a rigorous growth theory lens. It explicitly models the tension between AI automating learning tasks and augmenting human capital, offering theoretical insights into the long-term distributional consequences and structural shifts in labor demand relevant to the researcher's inquiry. How does the interaction between AI and human capital accumulation affect economic growth? I study this question in an endogenous growth model where human capital accumulates through learning-by-doing. AI augments human learning but also automates the tasks through which learning occurs. Since AI itself needs human capital supervision, the ironies appear: automation erodes the very skills that AI development demands. I show that if AI augments learning only modestly, the economy converges to a constant ratio of human to AI capital and endogenous growth stops. At the critical threshold, constant returns sustain endogenous balanced growth. Beyond it, the economy either explodes or collapses...	Enric Martorell
8	2026	Navigating Rule Ambiguity: Mapping the Human-AI Creativity Frontier ↗ This paper closely aligns with the project's core themes by providing a task-based framework to measure AI exposure, specifically distinguishing between human and machine creativity across occupations. It directly addresses the question of whether AI augments or substitutes for labor, offering empirical evidence on task reorganization and the boundaries of AI immunity in creative sectors. This paper empirically examines the boundary between human and machine creativity in cultural and creative industries by analysing 593 tasks across 126 creative occupations. We conceptualise creativity through an evolutionary perspective, viewing it as navigating rule ambiguity across three phases: retention (codified rules), adoption (tacit rules), and origination (novel rules). Employing synthetic data generation to measure cognitive and behavioural rule characteristics at task level, based on the December 2023 Australian Skills Classification, we identify two key mechanisms determining the creativity frontier. The structured novelty effect shows that AI autonomy increases significantly...	Emmanuelle Walkowiak
8	2025	GENERATIVE AI AND THE UNEVEN DISRUPTION OF FREELANCE LABOR: EXPOSURE, DIFFUSION, AND ADAPTATION UNDER CONSTRAINT ↗ [Title only] This paper directly addresses the core themes of AI's impact on online labor markets and distributional effects, specifically focusing on the uneven disruption faced by freelance workers. By examining exposure, diffusion, and adaptation under constraints, it provides critical evidence on how generative AI alters task reorganization and employment dynamics in the gig economy. No abstract available.	Željko Tekić, Arina Dunaeva

AI Impact on Jobs, Workers, and Wages

Core Questions

Key Findings

Identification & Credibility

Methods

Open Questions & Gaps

New additions — last 7 days

Full library (score ≥ 8 or seed)