|
on Artificial Intelligence |
| By: | Sukjin Han |
| Abstract: | The instrumental variables (IVs) method is a leading empirical strategy for causal inference. Finding IVs is a heuristic and creative process, and justifying its validity — especially exclusion restrictions — is largely rhetorical. We propose using large language models (LLMs) to search for new IVs through narratives and counterfactual reasoning, similar to how a human researcher would. The stark difference, however, is that LLMs can dramatically accelerate this process and explore an extremely large search space. We demonstrate how to construct prompts to search for potentially valid IVs. We contend that multi-step and role-playing prompting strategies are effective for simulating the endogenous decision-making processes of economic agents or social actors and for navigating language models through the realm of real-world scenarios. We apply our method to four well-known examples in economics: returns to schooling, demand and supply, and peer effects. Expert surveys reveal that some of the discovered IVs in each domain appear both novel and likely valid. |
| Date: | 2026–01–30 |
| URL: | https://d.repec.org/n?u=RePEc:bri:uobdis:26/833 |
| By: | Joel Persson; M{\aa}rten Schultzberg; Sebastian Ankargren |
| Abstract: | Organizations and researchers show increasing interest in using large language models (LLMs) in place of human participants in A/B tests, in the hope of experimenting faster and at lower cost. We study when a treatment effect estimated on LLM outcomes recovers the effect that would have been measured on the human population of interest. Distributional equivalence between LLM and human outcomes would make any standard estimator valid but is unrealistic. We therefore develop a statistical framework that adapts surrogate endpoint theory to LLMs. The framework shows that calibrating LLM outcomes to human outcomes identifies the average treatment effect under surrogacy and comparability conditions that are jointly weaker than distributional equivalence. When these conditions fail, the effect of interest is only partially identified, and we provide diagnostics that can falsify surrogacy on historical experiments together with a bound on the worst-case bias from limited overlap. We further show that the stochasticity inherent to LLMs introduces both bias and variance, but using an average of multiple draws as the surrogate mitigates both. We illustrate the methods and theory in simulations and an application to A/B tests on Upworthy headlines. A central takeaway from our work is that the validity of LLM outcomes as surrogates can only be falsified for past treatments and never verified for new ones, so human experiments remain indispensable for novel interventions. We discuss the role of LLM choice, prompting, and temperature as design variables, and how to size human experiments for validation. |
| Date: | 2026–06 |
| URL: | https://d.repec.org/n?u=RePEc:arx:papers:2606.17165 |
| By: | Drew Johnston; David Holtz; Alex Martin Richmond; Christopher Ong; Prasanna Tambe; Aaron Chatterji |
| Abstract: | We analyze usage data from OpenAI's Codex tool to present large-scale evidence of how agentic AI technology, which can take actions on a user's behalf, changes how people work. We use an automated, privacy-protecting pipeline to contrast usage across three populations: external personal-account users, external organizational-account users, and workers within OpenAI. We find that agentic AI usage is growing rapidly: the number of active users has grown more than fivefold in the first half of 2026, with the most rapid increase occurring outside the initial audience of software developers. Uptake is uneven: within OpenAI, Codex usage is nearly universal and has largely replaced business usage of ChatGPT. We document a similar shift to agentic tooling outside OpenAI, particularly within organizations, although external adoption remains lower and more uneven. In addition to headline usage figures, we observe measures of sophistication, and find that a growing number of users have used Codex to change their workflows substantially. More than 10% of users manage three or more concurrent Codex agents at some point each week and that 26.6% use skills, which allow users to share instructions for complex workflows. Alongside these changes in usage practices, request complexity has increased: since the start of the year, the share of individual Codex users who submit at least one request for a task estimated to require more than eight hours for an experienced human to complete has increased nearly tenfold. Concurrently, output has grown rapidly -- in June 2026, the median OpenAI employee in a legal role generated 13 times more monthly output tokens across Codex and ChatGPT than they did in November 2025, while the median researcher generated more than 50 times as many. We conclude by discussing the implications of these patterns for productivity, job reorganization, and workforce restructuring. |
| Date: | 2026–06 |
| URL: | https://d.repec.org/n?u=RePEc:arx:papers:2606.26959 |
| By: | ALEKSENKO, STEPAN; Miklos-Thal, Jeanine |
| Abstract: | We analyze the effects of information sharing in oligopoly when firms outsource pricing to a common third-party pricing algorithm developer. In a model where algorithms tailor prices to high-frequency demand shocks, we compare regimes that allow or prohibit conditioning on rival-specific shocks. Information sharing makes algorithmic prices more sensitive to seller-specific demand shocks---own and rival---and more correlated across sellers. These effects are stronger under common third-party algorithm design than under independent design because the third party's objective generates greater strategic complementarity in pricing than independent profit maximization. Information sharing harms expected consumer surplus more under common third-party design than under independent design, and its welfare effects are reversed across the two cases: information sharing lowers expected welfare under common third-party design while raising it under independent design. Our findings provide theoretical support for recent antitrust scrutiny of common third-party pricing algorithms that incorporate competitor data. |
| Keywords: | Algorithmic pricing; Information sharing; Antitrust; Oligopoly; Third-party sharing |
| JEL: | L13 L41 L42 D43 L86 |
| Date: | 2026–05 |
| URL: | https://d.repec.org/n?u=RePEc:cpr:ceprdp:21452 |
| By: | Conlon, John; Schwardmann, Peter |
| Abstract: | We examine whether sycophantic AI advice distorts decisions. Our experiment involves 1, 500 participants in 30 decision environments spanning core domains in economics and the social sciences. Contrary to the vast majority of predictions in an expert survey we conduct, we find that AI advice depolarizes choices on average, moving participants away from their initial leanings. This depolarization arises despite the LLM being measurably sycophantic: it disproportionately offers considerations that support users’ initial leanings and uses agreeable and flattering language. Depolarization occurs across moral and non-moral, objective and subjective, strategic and non-strategic, and complex and simple tasks. Increasing sycophancy weakens depolarization, showing that sycophancy is behaviorally relevant, even if it is generally outweighed by the informativeness of AI advice. Finally, several results mitigate the concern that market forces will generate greater polarizing effects outside the experiment or in the future. On the supply side, our baseline AI’s level of sycophancy is typical of leading models, and these models are not becoming more sycophantic over time. On the demand side, participants do not prefer greater sycophancy, do not select into AI advice in tasks where it is more polarizing, and exhibit greater depolarizing effects when they are more frequent AI users outside the experiment. |
| Keywords: | Large Language Models; Advice |
| JEL: | D83 O33 |
| Date: | 2026–05 |
| URL: | https://d.repec.org/n?u=RePEc:cpr:ceprdp:21505 |
| By: | Wayne Gao; Sukjin Han; Annie Liang |
| Abstract: | Large language models (LLMs) are increasingly used to predict human behavior. We propose a measure for evaluating how much knowledge a pretrained LLM brings to such a prediction: its equivalent sample size, defined as the amount of task-specific data needed to match the predictive accuracy of the LLM. We estimate this measure by comparing the prediction error of a fixed LLM in a given domain to that of flexible machine learning models trained on increasing samples of domain-specific data. We further provide a statistical inference procedure by developing a new asymptotic theory for cross-validated prediction error. Finally, we apply this method to the Panel Study of Income Dynamics. We find that LLMs encode considerable predictive information for some economic variables but much less for others, suggesting that their value as substitutes for domain-specific data differs markedly across settings. |
| Date: | 2026–01–30 |
| URL: | https://d.repec.org/n?u=RePEc:bri:uobdis:26/835 |
| By: | Cruces, Guillermo; Fernandez Meijide, Diego; Galiani, Sebastian; Galvez, Ramiro; Lombardi, María |
| Abstract: | Does generative artificial intelligence (AI) reinforce or reduce productivity differences across workers? Existing evidence largely studies AI within firms and occupations, where organizational selection compresses educational heterogeneity, leaving unclear whether AI narrows productivity gaps across individuals with substantially different levels of formal education. We address this question using a randomized online experiment conducted outside firms, in which 1, 174 adults ages 25–45 with heterogeneous educational backgrounds complete an incentivized, workplace-style business problem-solving task. The task is a general (not domain specific) exercise, and participants perform it either with or without access to a generative-AI assistant. Unlike prior work that studies heterogeneity within relatively homogeneous worker samples, our design targets the between–education-group productivity gap as the primary estimand. We find that AI increases productivity for all participants, with substantially larger gains for lower-education individuals. In the absence of AI access, higher-education participants outperform lower-education participants by 0.548 standard deviations; with AI access, this gap falls to 0.139 standard deviations, implying that generative AI closes about three quarters of the initial productivity gap. We interpret this pattern as evidence that generative AI narrows effective productivity differences in task execution by relaxing cognitive constraints that are more binding for lower-education individuals, even though underlying skill differences remain, as reflected in persistent education gaps in task performance and in a follow-up exercise without AI assistance. |
| Keywords: | Productivity; Inequality |
| JEL: | J24 O33 |
| Date: | 2026–03 |
| URL: | https://d.repec.org/n?u=RePEc:cpr:ceprdp:21299 |
| By: | Jason Abaluck; Kevin Bryan; Rebecca Ceppas de Castro; Basil Halperin; Todd Jones; Ezra Karger; Otto Kuusela; Dan Mayland; Ananaya Mittal; Connacher Murphy; Matt Reynolds; Josh Rosenberg; Philip Tetlock; Phil Trammell; Ria Viswanathan |
| Abstract: | We elicit forecasts of how AI will affect the U.S. economy, comparing the beliefs of five groups: academic economists, employees at AI companies, policy researchers focused on AI, highly accurate forecasters, and the general public. The median respondent in each group expects substantial advances in AI capabilities by 2030, small declines in labor force participation consistent with demographic shifts, and an annual GDP growth rate of 2.5%, which exceeds both the typical medium-run (2.0%) and long-run (1.7%) baseline forecasts from government agencies and private-sector forecasters. Conditional on a “rapid” AI progress scenario, in which AI systems surpass human performance on many cognitive and physical tasks, experts forecast substantial, though not historically unprecedented, economic shifts: annualized GDP growth rising to around 4% and the labor force participation rate falling from its current level of 62% to 55% by 2050, with roughly half of that decline—equivalent to around 10 million lost jobs—attributable to AI. A variance decomposition suggests that expert disagreement about these effects is driven primarily by different beliefs about the economic effects of highly capable AI systems rather than by disagreement about the pace of AI progress. These forecasts map onto notably different policy preferences across groups: experts strongly favor targeted measures such as worker retraining, whereas the general public supports both targeted programs and broader interventions, including a job guarantee and universal basic income. |
| Keywords: | Artificial intelligence; Economic forecasting; macroeconomic impacts |
| JEL: | O33 O38 O40 E27 J21 |
| Date: | 2026–03 |
| URL: | https://d.repec.org/n?u=RePEc:fip:fedhwp:103454 |
| By: | Baslandze, Salomé; Edwards, Zachary; Graham, John; McClure, Ty; Sparks, Michael; Meyer, Brent; Waddell, Sonya; Weitz, Daniel |
| Abstract: | 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. We document a productivity paradox, in which perceived productivity gains are larger than measured productivity gains, likely reflecting a delay in revenue realizations. In labor markets, we find little evidence of near-term aggregate employment declines due to AI, though larger companies anticipate AI-driven workforce reductions, while smaller firms expect modest gains. We also find evidence of compositional reallocation of labor both within and across firms, with routine clerical roles declining and a relative demand for skilled technical roles increasing. We develop an index that ranks job functions most negatively affected by AI. |
| Keywords: | Artificial intelligence; Productivity; Technological change; Labor markets; Occupations |
| JEL: | O33 D22 J24 |
| Date: | 2026–03 |
| URL: | https://d.repec.org/n?u=RePEc:cpr:ceprdp:21313 |
| By: | Ara Kharazian (Ramp); Lisa Simon (Revelio Labs); Ryan Stevens (Ramp) |
| Abstract: | We study how employment changes when firms adopt generative AI using observed AI spending from Ramp card and bill pay data linked to Revelio Labs workforce records for 21, 559 firms in the United States. We find that companies that adopt AI tend to grow faster following adoption, but the relationship is driven almost entirely by high-intensity adopters. Firms making the largest AI investments grow employment by roughly 10% following adoption, while low-intensity adopters see no statistically significant change. Entry-level headcount rises 12% for high-intensity adopters. Gains emerge gradually and are broad across roles, including engineering, sales, administration, and customer service. They are also uneven: adopters are already larger, more technical, faster-growing firms, and sector-level gains are concentrated in Information. The results counter predictions that AI adoption will lead to broad job loss. |
| Keywords: | artificial intelligence, generative AI, employment, labor markets, firm-level adoption, workforce adjustment |
| Date: | 2026–06–30 |
| URL: | https://d.repec.org/n?u=RePEc:epv:wpaper:ramp-ai-jobs-2026 |
| By: | Anabela Santos; Francesco Molica |
| Abstract: | Using novel data from the Survey on Access to Finance of Enterprises, whose wave conducted in the last quarter of 2025 introduced a dedicated module on AI adoption, and an Inverse Probability Weighted Regression Adjustment approach, we assess the conditional effect of artificial intelligence (AI) use by Small and Medium-sized Enterprises across twelve EU countries on the likelihood of employment growth. We find that AI adoption is associated with a positive conditional effect on the probability of employment growth, with no significant effect on the likelihood of employment decline, consistent with complementarity dominating substitution at the current stage of diffusionin Europe. Such employment gains scale with the depth of AI integration, and firms reporting significant AI use exhibit conditional effects twice as large as infrequent users, suggesting that the returns to AI depend on operational embeddedness rather than adoption per se. The estimated conditional effects are also strongly heterogeneous. Positive responses are concentrated among small and medium firms and in the services sector, while micro firms show no significant effect, and the construction sector is the only segment associated with an elevated risk of workforce contraction. Institutional context also matters decisively. Employment-enhancing effects are significantly larger in countries with stronger innovation ecosystems, more flexible labour markets, higher governance quality, and more decentralised political systems. These findings carry direct policy implications, promoting AI adoption without supporting deep integration, targeting micro enterprises, and accounting for national institutional capacity is unlikely to maximise the labour market gains from AI diffusion across Europe. |
| Keywords: | Employment growth; Small and Medium Enterprises; European Union |
| Date: | 2026–05 |
| URL: | https://d.repec.org/n?u=RePEc:ict:wpaper:2013/407834 |
| By: | Garicano, Luis; Li, Jin; Wu, Yanhui |
| Abstract: | This paper studies how the effect of AI on an occupation depends not just on which tasks AI can perform but also on how costly it is to unbundle those tasks from the job. Much of the discussion of AI and labor markets starts from task exposure: if AI can perform more tasks in an occupation, that occupation should lose employment or earnings. This is incomplete because labor markets price jobs, not tasks. Jobs bundle tasks together, and the effect of AI depends on how costly it is to break the bundle. We build a two-task model in which AI can either assist one task inside a bundled job or supply that task autonomously while a human supplies the residual task. We show that, in weak-bundle occupations, AI automates some tasks and narrows the boundary of the job, activating the standard task-substitution channel once product demand is sufficiently inelastic. In strong-bundle occupations where tasks are not independently reallocable, AI improves performance inside the job, but does not remove the human from the bundle. Thus, bundling provides a force that protects jobs and workers' share of downstream revenue. |
| JEL: | J24 J23 O33 L23 D20 J31 |
| Date: | 2026–05 |
| URL: | https://d.repec.org/n?u=RePEc:cpr:ceprdp:21453 |
| By: | Yanyou Chen; Mitchell Hoffman; Huilan Xu; Zhe Yuan |
| Abstract: | Can AI improve workplace outcomes for workers with disabilities? We examine the relative performance of deaf or hard of hearing (DHH) workers on one of China's largest food-delivery platforms. Pre-AI, DHH workers are slower than non-disabled workers and have worse customer ratings, although they supply more hours to the platform and are less likely to quit. Midway through our data, the platform suddenly introduces an AI-based intelligent outbound calling system designed to improve customer communication for DHH workers. Using a difference-in-differences design comparing DHH and non-disabled workers before and after the AI tool, we find that AI increases the speed and productivity of DHH workers, especially on tasks involving customer interaction; substantially reduces negative customer ratings; and increases labor supply and retention. AI eliminates one-third of the disability hourly pay gap and significantly increases the profitability of DHH workers for the platform. |
| JEL: | J14 M50 |
| Date: | 2026–06 |
| URL: | https://d.repec.org/n?u=RePEc:nbr:nberwo:35372 |
| By: | Bick, Alexander; Blandin, Adam; Deming, David; Fuchs-Schündeln, Nicola; Jessen, Jonas |
| Abstract: | This paper combines international evidence from worker and firm surveys conducted in 2025 and 2026 to document large gaps in AI adoption, both between the US and Europe and across European countries. Cross-country differences in worker demographics and firm composition account for an important share of these gaps. AI adoption, within and across countries, is also closely linked to firm personnel management practices and whether firms actively encourage AI use by workers. Micro-level evidence suggests that AI generates meaningful time savings for many workers. At the macro level, in recent years industries with higher AI adoption rates have experienced faster productivity growth. While we do not establish causality, this relationship is statistically significant and similar in magnitude in Europe and the US. We find no clear evidence that industry-level AI adoption is associated with employment changes. We discuss limitations of existing data and outline priorities for future data collection to better assess the productivity and labor market effects of AI. |
| Keywords: | Generative AI |
| JEL: | J24 M16 O14 O33 |
| Date: | 2026–03 |
| URL: | https://d.repec.org/n?u=RePEc:cpr:ceprdp:21337 |
| By: | Bilgin, Nuriye Melisa; Ottaviano, Gianmarco |
| Abstract: | We study how digital infrastructure relaxes constraints on the diffusion and economic impact of artificial intelligence (AI). Using administrative data and a nationally representative enterprise survey from Turkey (2021–2024), we document significant disparities in AI adoption. Adoption is concentrated among large firms and in regions with high-speed broadband and proximity to data centers, particularly for software-intensive and cloud-based applications. To identify causal effects, we exploit the staggered expansion of Turkey’s national natural gas pipeline network, which serves as a conduit for fiber-optic deployment. Because pipeline routing is determined by energy distribution priorities rather than digital demand, it provides plausibly exogenous variation in connectivity. Difference-in-differences estimates show that improved connectivity significantly increases AI adoption, particularly for software-intensive technologies and among small and medium-sized enterprises. Instrumental-variable estimates indicate that infrastructure-driven AI adoption raises labor productivity and export intensity while shifting labor composition toward ICT-related roles. These findings highlight digital infrastructure as a primary determinant of both the pace of AI diffusion and its resulting economic returns. |
| Keywords: | Artificial intelligence; Digital infrastructure; Broadband; Technology diffusion; Firm productivity; Cloud computing |
| JEL: | O33 L86 D24 J24 O14 R12 |
| Date: | 2026–04 |
| URL: | https://d.repec.org/n?u=RePEc:cpr:ceprdp:21385 |
| By: | Hankui Wang (Université Côte d'Azur, CNRS, GREDEG, France); Jiachen Yi (School of Economics, University of Bristol, United Kingdom); Philipp Harting (Université Côte d'Azur, CNRS, GREDEG, France) |
| Abstract: | Artificial Intelligence (AI) is adopted far faster by individuals than by organisations, yet most diffusion models treat both populations as homogeneous and independent. We develop an extended Bass model with heterogeneous archetypes, bidirectional cross-group spillovers, endogenous barrier decay, and productivity feedback. Base parameters are estimated from survey data on individual and organisational AI adoption across 45 countries (2020-2025); the extended model is calibrated and simulated over ten years. Two findings emerge: cross-group spillover from individual to organisational adoption is the dominant diffusion accelerator in the cross-country evidence, while within-group imitation is undetectable in the current two-year panel; and heterogeneity generates a 5.2-year spread across firm archetypes in time to 50% adoption, capturing most of the 6-year empirical firm-size gap. The calibrated model simulates overall organisational AI adoption reaching 50% by 2029 and 74% by 2033 under the benchmark calibration, translating into sustained productivity growth. Education and training delivers the largest adoption gain per unit of intervention intensity, while direct subsidies most effectively narrow the adoption gap between large and small firms. |
| Keywords: | artificial intelligence; technology diffusion; Bass model |
| JEL: | O33 O31 O38 O53 O14 |
| Date: | 2026–06 |
| URL: | https://d.repec.org/n?u=RePEc:gre:wpaper:2026-16 |
| By: | Christophe André; Matthias Schief |
| Abstract: | Demographic headwinds are set to weaken economic growth in OECD countries over the coming decades. At the same time, artificial intelligence (AI) provides opportunities for productivity gains, potentially alleviating labour shortages and boosting economic growth. However, little is known about how exposure to AI varies over the life cycle and what this may imply for AI deployment in ageing societies. This paper shows, using OECD Programme for the International Assessment of Adult Competencies (PIAAC) data, that workers’ overall exposure to AI (automation and augmentation) exhibits an inverted U-shaped pattern across age groups, albeit less pronounced when controlling for education, occupation and country. Exposure to automation is higher in younger age groups and declines rapidly with age, as experience tends to complement AI. Nevertheless, as a general-purpose technology, AI is bound to be disruptive. Reaping its benefits will require labour market reallocation, reskilling and upskilling, and business dynamism and innovation, which may all be weaker in ageing societies. |
| Keywords: | ageing, artificial intelligence, business dynamism, demography, economic growth, innovation, labour market policies, lifelong learning, PIAAC, productivity, reskilling, technological change, upskilling |
| JEL: | J08 J11 J24 O33 O40 |
| Date: | 2026–07–06 |
| URL: | https://d.repec.org/n?u=RePEc:oec:ecoaaa:1870-en |
| By: | Bilgin, Nuriye Melisa; Ottaviano, Gianmarco |
| Abstract: | Do the determinants of technology adoption depend on technological architecture? Using administrative data on Turkish firms from 2021 to 2024, we compare the adoption of traditional and generative artificial intelligence (GenAI). We show that GenAI adoption is driven by workforce skill intensity and is not positively associated with firm size, whereas traditional AI depends on both scale and skills. Firms that adopt both technologies are distinct and represent the most persistent adoption mode. Conditional on adoption, the skill-to-size ratio governs technology choice, and transition dynamics indicate a sequential process in which firms adopt GenAI before expanding to hybrid use. Exploiting the release of ChatGPT as a quasi-experimental reduction in access costs, we find that high-skill firms differentially increased GenAI adoption, while firm size played a limited role. These results suggest that the canonical size-based diffusion pattern is not universal but depends on the cost structure of technologies, with implications for innovation policy and productivity dispersion. |
| Keywords: | Artificial intelligence |
| JEL: | O33 L25 D22 O14 J3 |
| Date: | 2026–05 |
| URL: | https://d.repec.org/n?u=RePEc:cpr:ceprdp:21506 |
| By: | Döttling, Robin; Emery, Logan P.; Zhao, Shuo |
| Abstract: | Generative AI has the potential to transform corporate innovation, but intellectual property (IP) created without sufficient human input is ineligible for protection by IP systems. We model a firm's choice of AI versus human-capital use when investing in innovation, with IP protection granted based on a noisy signal of human-capital use. We derive the IP policy's effect on incentives and characterize when the IP system can "kill" AI use. Alternatively, low AI costs can "kill" the IP system or shift its role to providing a human-capital subsidy, depending on signal noise and the social value of human-capital use in innovation. When consumers value human-created works, human-capital use is distorted by an adverse selection discount. The IP policy can mitigate this by deterring high-cost firms' investment, or by acting as a credible signal of incentives for human-capital use that triggers a positive feedback loop through consumer beliefs. |
| Keywords: | Generative AI; Innovation; Copyright; Intellectual property protection; Adverse selection |
| JEL: | G31 G38 O31 O34 O38 |
| Date: | 2026–04 |
| URL: | https://d.repec.org/n?u=RePEc:cpr:ceprdp:21399 |
| By: | Masayuki Morikawa |
| Abstract: | This study documents the adoption of AI in the workplace and its impact on productivity among workers in the Japanese machinery industry. At the end of 2025, 34% of workers use AI in their jobs, with R&D accounting for the largest proportion of AI-utilized jobs. Among AI users, the mean share of tasks using AI, efficiency gains, and resulting productivity effects are 12%, 20%, and 4%, respectively. Most workers use AI for only a small fraction of their overall job tasks. The productivity effect is larger for continuous AI users than for new AI users, suggesting selection and learning effects of AI adoption. The use of AI at work is projected to increase labor productivity in the industry by 0.3-0.4 percentage points annually over the next several years. If the use of AI in R&D activities improves the efficiency of R&D investment, it is likely to generate productivity gains that extend beyond simple labor-saving effects. Finally, more than 80% of workers hold positive views toward expanding the use of AI in the workplace, with stronger support among those already using AI and those facing severe labor shortages. |
| Keywords: | artificial intelligence, machinery industry, productivity |
| JEL: | J24 L60 O33 |
| Date: | 2026–07 |
| URL: | https://d.repec.org/n?u=RePEc:een:camaaa:2026-49 |
| By: | Chau Tran Bao; Khoi Nguyen Dinh Nguyen; Ha Nguyen Manh; Ngan Nguyen Thi Thuy |
| Abstract: | Automation and artificial intelligence (AI) are reshaping labor demand unevenly across space, creating an urgent imperative for place-sensitive education and workforce policy. This study asks whether regional exposure to automation and to AI relates to local employment and wages in opposite ways, and whether those relationships differ between urban and rural regions -- two questions whose answers carry direct implications for how skills training and digital education should be targeted. Using a region-by-year panel and shift-share measures of technological exposure built from baseline industry and occupation composition, we estimate two-way fixed-effects and instrumental-variable models that interact exposure with an urban indicator. The framework distinguishes automation exposure, concentrated in routine work, from AI exposure, concentrated in cognitive work -- a distinction that maps directly onto the types of skills that education systems need to develop or preserve. Estimates show automation exposure lowering employment and wages, with the employment loss cushioned in cities, while AI exposure raises wages and concentrates in urban regions. Technology therefore reshapes, rather than simply widens, the divide. The findings argue for place-sensitive policy: weighting reallocation and reskilling support toward routine-exposed rural regions, while extending digital infrastructure and AI-complementary skills outward so that rural workers can share AI's wage gains rather than absorb only automation's losses. |
| Date: | 2026–06 |
| URL: | https://d.repec.org/n?u=RePEc:arx:papers:2606.22833 |
| By: | Lenzu, Simone |
| Abstract: | I develop a framework analyzing how artificial intelligence (AI) reshapes monetary policy through three interrelated channels: cyclical transmission, structural transition, and financial stability. In the short run, AI can alter inflation dynamics by changing how supply and demand disturbances map into prices — through shifts in production technologies, pricing behavior, cost pass-through, and expectations — even when conventional measures of economic slack are unchanged. Over longer horizons, AI may shift the natural benchmarks around which policy is calibrated, including potential output and the natural rate of interest. For financial stability, AI may improve credit allocation and risk assessment, but can also heighten systemic vulnerabilities through inflated expectation-driven asset valuations and model monocultures. A particular risk arises at the intersection of these channels: if AI initially depresses realized efficiency through adoption frictions while simultaneously fueling elevated asset valuations, the economy may face cost-push inflation and financial fragility at once — an AI-specific stagflation risk that the interest rate instrument alone is ill-suited to address. I argue that AI does not call for a redefinition of central banks' objectives, but it does require a recalibration of existing frameworks: its diffusion blurs the distinction between cyclical fluctuations and structural shifts, raising the value of cost-side diagnostics and robust policy strategies over exclusive reliance on reduced-form inflation-gap relationships. |
| JEL: | O33 E52 E58 E31 E32 E44 |
| Date: | 2026–03 |
| URL: | https://d.repec.org/n?u=RePEc:cpr:ceprdp:21248 |
| By: | Rerotlhe B. Basele (Macquarie University); Peter C.B. Phillips (Yale University, University of Auckland, Singapore Management University); Shuping Shi (Macquarie University) |
| Abstract: | The AI boom has driven the Nasdaq and the Magnificent Seven tech stocks to record highs. But how much do these new records reflect underlying value, how much is speculation, and how vulnerable are these stocks and the wider market to a major downturn? Our evidence and analyses show clear signs of bubble exuberance in most of these stocks, concentrated in a few names like Nvidia, leading to latent risks for investors who assume their index funds are safely diversified and supported by wider economic fundamentals. |
| Date: | 2026–06 |
| URL: | https://d.repec.org/n?u=RePEc:cwl:cwldpp:2537 |
| By: | Fornaro, Luca; Wolf, Martin |
| Abstract: | We provide a macroeconomic framework to study monetary and fiscal policies for AI. Advances in AI expand firms' ability to automate production. While higher automation boosts productivity and potential output, it also reduces workers' share of income. Since workers have a high propensity to consume, advances in AI may depress aggregate demand and lead to a slump. Expansionary monetary policy can convert an AI slump into an AI boom, but in doing so it faces two challenges. In the short run, AI worsens the inflation-employment trade off faced by the central bank. In the medium run, monetary policy may be constrained by the zero lower bound, since weak demand lowers the natural rate. Employment subsidies and cuts in labor taxes can usefully complement monetary policy, by reducing firms' cost of labor and inflation, as well as supporting workers' income and aggregate demand. |
| Keywords: | Inflation |
| JEL: | E32 E43 E52 O31 O42 |
| Date: | 2026–04 |
| URL: | https://d.repec.org/n?u=RePEc:cpr:ceprdp:21412 |
| By: | Wang Jin; Nadav Kunievsky; Bowen Lou; Tianshu Sun; James Evans |
| Abstract: | Over the past decade, U.S. policies have increasingly aimed to preserve artificial intelligence (AI) leadership by promoting domestic free-market policies while controlling global technological chokepoints, particularly advanced semiconductors and computational infrastructure. These measures raised the cost of Chinese AI development, but they also increased the strategic value of open and locally adaptable AI systems. Before raising export controls on high-performance chips, both the U.S. and China promoted policies that included support for open-source AI. During the period following major U.S. export-control shocks, China increasingly embedded open-source AI into national technology strategy through proposed ecosystem building, standards coordination, and resilience-oriented deployment. Moreover, Chinese developers increased engagement with open-source large language model repositories substantially more than U.S. developers did, consistent with a shift toward open infrastructure under geopolitical constraints. Subsequently, Chinese-origin open models diffused widely through open-source communities and scientific research. Even though such models remained largely absent from U.S. patent disclosures, American commercial entities use them in open-access research, suggesting their undermeasured importance within the foundation of U.S. commercial activity. These findings suggest that technological containment policies may unintentionally accelerate open innovation ecosystems as a competitive response, with implications for global leadership in both academic and commercial artificial intelligence. |
| Date: | 2026–06 |
| URL: | https://d.repec.org/n?u=RePEc:arx:papers:2606.15999 |
| By: | Lasse Dierich; Orestis Schinas |
| Abstract: | Ship finance is a data-intensive and document-heavy segment of asset-based lending, requiring the integration of financial, technical, contractual, and regulatory information from heterogeneous and largely unstructured sources. Increasing environmental regulation and ESG reporting requirements are adding further complexity to underwriting and loan-origination processes. Recent advances in artificial intelligence (AI), particularly large language models (LLMs), create new opportunities for processing and analysing such information. This paper reviews potential applications of AI in ship finance, with a particular focus on LLM-based systems for document comprehension, information extraction, and workflow automation. We present ShipFinance.ai, a modular agentic architecture to support loan application workflows in ship finance. The proposed system combines an LLM-based extraction module, financial analysis components, external maritime data services, and a controlled document-generation module with a chatbot interface to support the preparation of standardized financing applications. The paper discusses the key challenges for using such models in production. We argue that AI-assisted systems can support maritime finance professionals in managing increasingly complex information and reporting requirements. |
| Date: | 2026–05 |
| URL: | https://d.repec.org/n?u=RePEc:arx:papers:2606.11238 |