|
on Artificial Intelligence |
| By: | Erik Brynjolfsson; Anton Korinek; Ajay K. Agrawal |
| Abstract: | As we approach Transformative Artificial Intelligence (TAI), there is an urgent need to advance our understanding of how it could reshape our economic models, institutions and policies. We propose a research agenda for the economics of TAI by identifying nine Grand Challenges: economic growth, innovation, income distribution, decision-making power, geoeconomics, information flows, safety risks, human well-being, and transition dynamics. By accelerating work in these areas, researchers can develop insights and tools to help fulfill the economic potential of TAI. |
| JEL: | A11 O33 O40 |
| Date: | 2025–09 |
| URL: | https://d.repec.org/n?u=RePEc:nbr:nberwo:34256 |
| By: | Frédéric Marty; Thierry Warin |
| Abstract: | The rapid advancement of generative artificial intelligence (AI) is increasingly shaped by control over two critical inputs: high-quality data and the compute infrastructure required to train and update large-scale model weights. This paper argues that these inputs – rather than algorithmic talent or novel architectures alone – have become the decisive strategic assets in generative AI, creating steep structural barriers to entry. We examine who controls these resources and how this control is territorially distributed across countries. Building on literature in industrial organization, competition policy, and international political economy, we highlight a gap in existing research: insufficient attention to the territorial concentration of “model-weight-setting” capacity, i.e. the ability to train cutting-edge foundation models. We find that the capacity to set foundation model weights is overwhelmingly concentrated in a few firms and regions, reinforcing market concentration and limiting the AI development sovereignty of most countries. While innovations in model architectures and efficiency (illustrated by the DeepSeek case) can reduce compute requirements at the margin, they do not eliminate the scale advantages conferred by privileged access to massive proprietary datasets and nation-scale computing clusters. The paper concludes with implications for competition and regulation, arguing that the territorial control of data and compute resources is a fundamental structural challenge for both market competition and global equity in AI. Les progrès rapides de l’intelligence artificielle générative (IA) sont de plus en plus conditionnés par le contrôle de deux intrants essentiels : des données de haute qualité et l’infrastructure de calcul nécessaire pour entraîner et actualiser les poids de modèles à grande échelle. Cet article soutient que ces intrants – plutôt que le seul talent algorithmique ou la nouveauté des architectures – sont devenus les actifs stratégiques décisifs de l’IA générative, créant ainsi d’importantes barrières structurelles à l’entrée. Nous examinons qui contrôle ces ressources et comment ce contrôle se répartit territorialement entre les pays. En nous appuyant sur la littérature en organisation industrielle, en politique de concurrence et en économie politique internationale, nous mettons en évidence une lacune dans les recherches existantes : l’attention insuffisante portée à la concentration territoriale de la « capacité de réglage des poids des modèles », c’est-à-dire la faculté d’entraîner des modèles de fondation de pointe. Nos résultats montrent que cette capacité est largement concentrée dans quelques entreprises et régions, ce qui renforce la concentration des marchés et limite la souveraineté de la plupart des pays en matière de développement de l’IA. Bien que les innovations en matière d’architectures de modèles et d’efficacité (comme l’illustre le cas DeepSeek) puissent réduire les besoins en calcul à la marge, elles n’éliminent pas les avantages d’échelle conférés par l’accès privilégié à d’immenses ensembles de données propriétaires et à des grappes de calcul de dimension nationale. L’article conclut en soulignant les implications pour la concurrence et la régulation, en avançant que le contrôle territorial des données et des ressources de calcul constitue un défi structurel fondamental pour la concurrence sur les marchés et pour l’équité mondiale en matière d’IA. |
| Keywords: | Generative Artificial Intelligence, Data Sovereignty, Compute Infrastructure, Competition Policy, Territorial Concentration, Intelligence artificielle générative, Souveraineté des données, Infrastructure de calcul, Politique de concurrence, Concentration territoriale |
| Date: | 2025–09–19 |
| URL: | https://d.repec.org/n?u=RePEc:cir:cirwor:2025s-27 |
| By: | Xienan Cheng; Mustafa Dogan; Pinar Yildirim |
| Abstract: | This study investigates the effects of artificial intelligence (AI) adoption in organizations. We ask: (1) How should a principal optimally deploy limited AI resources to replace workers in a team? (2) In a sequential workflow, which workers face the highest risk of AI replacement? (3) How does substitution with AI affect both the replaced and non-replaced workers’ wages? We develop a sequential team production model in which a principal can use peer monitoring—where each worker observes the effort of their predecessor—to discipline team members. The principal may replace some workers with AI agents, whose actions are not subject to moral hazard. Our analysis yields four key results. First, the optimal AI strategy stochastically replaces workers rather than fixating on a single position. Second, the principal replaces workers at the beginning and at the end of the workflow, but does not replace the middle worker, since this worker is crucial for sustaining the flow of information obtained by peer monitoring. Third, the principal may optimally underutilize available AI capacity. Fourth, the optimal AI adoption increases average wages and reduces intra-team wage inequality. |
| JEL: | D20 L0 M0 M12 M2 M21 M5 M52 |
| Date: | 2025–09 |
| URL: | https://d.repec.org/n?u=RePEc:nbr:nberwo:34259 |
| By: | Isaiah Andrews; Maryam Farboodi |
| Abstract: | Economic theory predicts that transformative technologies may influence interest rates by changing growth expectations, increasing uncertainty about growth, or raising concerns about existential risk. Examining US bond yields around major AI model releases in 2023-4, we find economically large and statistically significant movements concentrated at longer maturities. The median and mean yield responses across releases in our sample are negative: long-term Treasury, TIPS, and corporate yields fall and remain lower for weeks. Viewed through the lens of a simple, representative agent consumption-based asset pricing model, these declines correspond to downward revisions in expected consumption growth and/or a reduction in the perceived probability of extreme outcomes such as existential risk or arrival of a post-scarcity economy. By contrast, changes in consumption growth uncertainty do not appear to drive our results |
| JEL: | E43 E44 G1 G12 G14 O30 O4 |
| Date: | 2025–09 |
| URL: | https://d.repec.org/n?u=RePEc:nbr:nberwo:34243 |
| By: | Yijia Xiao; Edward Sun; Tong Chen; Fang Wu; Di Luo; Wei Wang |
| Abstract: | Developing professional, structured reasoning on par with human financial analysts and traders remains a central challenge in AI for finance, where markets demand interpretability and trust. Traditional time-series models lack explainability, while LLMs face challenges in turning natural-language analysis into disciplined, executable trades. Although reasoning LLMs have advanced in step-by-step planning and verification, their application to risk-sensitive financial decisions is underexplored. We present Trading-R1, a financially-aware model that incorporates strategic thinking and planning for comprehensive thesis composition, facts-grounded analysis, and volatility-adjusted decision making. Trading-R1 aligns reasoning with trading principles through supervised fine-tuning and reinforcement learning with a three-stage easy-to-hard curriculum. Training uses Tauric-TR1-DB, a 100k-sample corpus spanning 18 months, 14 equities, and five heterogeneous financial data sources. Evaluated on six major equities and ETFs, Trading-R1 demonstrates improved risk-adjusted returns and lower drawdowns compared to both open-source and proprietary instruction-following models as well as reasoning models. The system generates structured, evidence-based investment theses that support disciplined and interpretable trading decisions. Trading-R1 Terminal will be released at https://github.com/TauricResearch/Tradin g-R1. |
| Date: | 2025–09 |
| URL: | https://d.repec.org/n?u=RePEc:arx:papers:2509.11420 |