|
on Artificial Intelligence |
| By: | Igor Sadoune; Marcelin Joanis; Andrea Lodi |
| Abstract: | This paper introduces the Minimum Price Markov Game (MPMG), a theoretical model that reasonably approximates real-world first-price markets following the minimum price rule, such as public auctions. The goal is to provide researchers and practitioners with a framework to study market fairness and regulation in both digitized and non-digitized public procurement processes, amid growing concerns about algorithmic collusion in online markets. Using multi-agent reinforcement learningdriven artificial agents, we demonstrate that (i) the MPMG is a reliable model for first-price market dynamics, (ii) the minimum price rule is generally resilient to non-engineered tacit coordination among rational actors, and (iii) when tacit coordination occurs, it relies heavily on self-reinforcing trends. These findings contribute to the ongoing debate about algorithmic pricing and its implications. Cet article présente le jeu du prix minimum de Markov (MPMG), un modèle théorique qui se rapproche raisonnablement des marchés réels qui suivent la règle du prix minimum, tels que les enchères publiques. L'objectif est de fournir aux chercheurs et aux praticiens un cadre pour étudier l'équité du marché et la réglementation dans les processus de marchés publics numériques et non numériques, dans un contexte de préoccupations croissantes concernant la collusion algorithmique sur les marchés en ligne. En utilisant des agents artificiels basés sur l'apprentissage par renforcement multi-agents, nous démontrons que (i) le MPMG est un modèle fiable pour la dynamique du marché au premier prix, (ii) la règle du prix minimum est généralement résistante à la coordination tacite non technique entre les acteurs rationnels, et (iii) lorsque la coordination tacite se produit, elle s'appuie fortement sur des tendances qui se renforcent d'elles-mêmes. Ces résultats contribuent au débat en cours sur la tarification algorithmique et ses implications. |
| Keywords: | Algorithmic Game Theory, Multiagent Reinforcement Learning, Algorithmic Coordination, Algorithmic Pricing, Théorie algorithmique des jeux, apprentissage par renforcement multi-agents, coordination algorithmique, tarification algorithmique |
| Date: | 2025–04–08 |
| URL: | https://d.repec.org/n?u=RePEc:cir:cirwor:2025s-07 |
| By: | Fontanelli, Luca; Guerini, Mattia; Miniaci, Raffaele; Secchi, Angelo |
| Abstract: | While artificial intelligence (AI) adoption holds the potential to enhance business operations through improved forecasting and automation, its relation with average productivity growth remain highly heterogeneous across firms. This paper shifts the focus and investigates the impact of predictive artificial intelligence (AI) on the volatility of firms’ productivity growth rates. Using firm-level data from the 2019 French ICT survey, we provide robust evidence that AI use is associated with increased volatility. This relationship persists across multiple robustness checks, including analyses addressing causality concerns. To propose a possible mechanisms underlying this effect, we compare firms that purchase AI from external providers (“AI buyers”) and those that develop AI in-house (“AI developers”). Our results show that heightened volatility is concentrated among AI buyers, whereas firms that develop AI internally experience no such effect. Finally, we find that AI-induced volatility among “AI buyers” is mitigated in firms with a higher share of ICT engineers and technicians, suggesting that AI’s successful integration requires complementary human capital. |
| Keywords: | Dairy Farming, Production Economics, Research and Development/Tech Change/Emerging Technologies, Resource/Energy Economics and Policy |
| Date: | 2025–04–07 |
| URL: | https://d.repec.org/n?u=RePEc:ags:feemwp:355806 |
| By: | Donatella Gatti (Sorbonne Paris Nord University, CEPN); Julien Vauday (Sorbonne Paris Nord University, CEPN) |
| Abstract: | This paper investigates the transmission of environmentalist values among citizens in relation to the diffusion of generative AI devices. This diffusion occurs following races in which two strategies, i.e. Safe AI and Unsafe AI, are available to firms and adopted according to an evolutionary game framework. The adoption of Unsafe AI leads to disasters according to a probability associated with both technological and environmental damage. The latter is influenced by consumption patterns and social change resulting from the adoption of environmentalist vs. materialist values by future citizens. A new socialization channel is proposed, which operates through the interactions between humans and AI devices as role models. The interplay between AI diffusion and social change results in complex dynamics leading to multiple steady states. Based on risk-dominance, we show that a complementarity between Safe AI and environmentalism might emerge only if the endogenous disaster probability is higher than a threshold p. |
| Keywords: | Environmental transition, artificial intelligence, AI races, social change, carbon tax, disasters |
| JEL: | O |
| Date: | 2025 |
| URL: | https://d.repec.org/n?u=RePEc:inf:wpaper:2025.5 |
| By: | Jaroslaw Kornowicz (Paderborn University) |
| Abstract: | This study empirically examines the "Evaluative AI" framework, which aims to enhance the decision-making process for AI users by transitioning from a recommendation-based approach to a hypothesis-driven one. Rather than offering direct recommendations, this framework presents users pro and con evidence for hypotheses to support more informed decisions. However, findings from the current behavioral experiment reveal no significant improvement in decision-making performance and limited user engagement with the evidence provided, resulting in cognitive processes similar to those observed in traditional AI systems. Despite these results, the framework still holds promise for further exploration in future research. |
| Keywords: | explainable AI, human-computer interaction, human-ai interaction, decision support system |
| JEL: | C91 D81 C88 O33 |
| Date: | 2025–03 |
| URL: | https://d.repec.org/n?u=RePEc:pdn:dispap:134 |
| By: | Tao Chen; Shuwen Pi; Qing Sophie Wang (University of Canterbury) |
| Abstract: | This study examines the impact of artificial intelligence (AI) on corporate investment efficiency. Our analysis of recruitment data from Chinese listed companies reveals a positive correlation between AI and investment efficiency, primarily driven by a reduction in over-investment. Specifically, a one-standard-deviation increase in AI hiring is associated with a 3.1% improvement in investment efficiency. This improvement results from better investment decisions (e.g., greater responsiveness to growth opportunities and fewer value-destroying mergers and acquisitions), and more effective internal capital allocation (e.g., improvements in innovation and operational efficiency). The positive impact of AI is stronger in firms with less government intervention, flatter organizational structures, technically experienced boards, poorer information environments, and traditional and lowly competitive industries. Overall, our findings highlight the importance of AI skills in shaping corporate investment decisions. |
| Keywords: | Artificial intelligence, AI hiring, investment efficiency |
| JEL: | O14 O33 G31 G34 |
| Date: | 2025–04–01 |
| URL: | https://d.repec.org/n?u=RePEc:cbt:econwp:25/05 |
| By: | Natalie Brose; Christian Spielmann; Christian Tode |
| Abstract: | Since the public release of ChatGPT in late 2022, the role of Generative AI chatbots in education has been widely debated. While some see their potential as automated tutors, others worry that inaccuracies and hallucinations could harm student learning. Thisstudy assesses ChatGPT models (GPT-3.5, GPT-4o, and o1preview) across important dimensions of student learning by evaluating their capabilities and limitations to serve as a non-interactive, automated tutor. In particular, we analyse performance in two tasks commonly used in principles of economics courses: explaining economic concepts and answering multiple-choice questions. Our findings indicate that newer models generate very accurate responses, although some inaccuracies persist. A key concern is that ChatGPT presents all responses with full confidence, making errors difficult for students to recognize. Furthermore, explanations are often quite narrow, lacking holistic perspectives, and the quality of examples remains poor. Despite these limitations, we argue that ChatGPT can serve as an effective automated tutor for basic, knowledge-based questions—supportingstudents while posing a relatively low risk of misinformation. Educators can hence recommend Generative AI chatbots for student learning, but should teach students the limitations of the technology. |
| Date: | 2025–04–02 |
| URL: | https://d.repec.org/n?u=RePEc:bri:uobdis:25/786 |