|
on Artificial Intelligence |
| By: | Edmundo Molina-Perez (School of Government and Public Transformation, Tecnológico de Monterrey); Pedro Cortes (Tecnológico de Monterrey); Isaac Molina (Tecnológico de Monterrey); Fernanda Sobrino (Tecnológico de Monterrey); Mario Tellez (Tecnológico de Monterrey); Yessica Orozco (Tecnológico de Monterrey); Mitzi Castellón (Tecnológico de Monterrey); Steven Popper (Tecnológico de Monterrey); Luis Serra (Tecnológico de Monterrey) |
| Abstract: | Decision-making is a multifaceted cognitive process influenced by task complexity, information availability, individual cognitive strategies, and environmental settings. Yet, the neural mechanisms guiding everyday choices remain incompletely understood. This gap intensifies when integrating real-time aids, such as artificial intelligence tools (AIT), as cognitive decisionsupport especially for complex and ambiguous problems. This study explores the neural mechanisms of decision-making and examines how AIT influences these processes. Combining behavioral assessments and neurophysiological measurements, we investigate the dynamic interplay between human cognition and AIT through behavioral execution and electroencephalogram (EEG) activity. Experimental data from 54 participants suggest that in low-complexity decision-making, AIT is largely ignored in favor of heuristics. In high-complexity contexts, AIT positively influences decision-making outcomes while also increasing capacity for engagement with a challenging task as registered by EEG cortical activity. This suggests a non-linear effect of AIT in decision-making strategies highlighting its role as a complement to —rather than a replacement of—human cognitive processes. |
| Keywords: | artificial intelligence, decision-making, EEG, neuroeconomics, cognitive support tools |
| JEL: | C91 D83 D89 |
| Date: | 2025–08 |
| URL: | https://d.repec.org/n?u=RePEc:gnt:wpaper:8 |
| By: | Victor Klockmann (JMU - Julius-Maximilians-Universität Würzburg = University of Würzburg [Würsburg, Germany], Goethe University Frankfurt = Goethe-Universität Frankfurt am Main, Max Planck Institute for Human Development - Max-Planck-Gesellschaft); Alicia von Schenk (JMU - Julius-Maximilians-Universität Würzburg = University of Würzburg [Würsburg, Germany], Goethe University Frankfurt = Goethe-Universität Frankfurt am Main, Max Planck Institute for Human Development - Max-Planck-Gesellschaft); Marie Claire Villeval (GATE Lyon Saint-Étienne - Groupe d'Analyse et de Théorie Economique Lyon - Saint-Etienne - UL2 - Université Lumière - Lyon 2 - UJM - Université Jean Monnet - Saint-Étienne - EM - EMLyon Business School - CNRS - Centre National de la Recherche Scientifique) |
| Abstract: | In the field of machine learning, the decisions of algorithms depend on extensive training data contributed by numerous, often human, sources. How does this property affect the social nature of human decisions that serve to train these algorithms? By experimentally manipulating the pivotality of individual decisions for a supervised machine learning algorithm, we show that the diffusion of responsibility weakened revealed social preferences, leading to algorithmic models favoring selfish decisions. Importantly, this phenomenon cannot be attributed to shifts in incentive structures or the presence of externalities. Rather, our results suggest that the expansive nature of Big Data fosters a sense of diminished responsibility and serves as an excuse for selfish behavior that impacts individuals and the whole society. |
| Keywords: | Artificial intelligence, Big data, Pivotality, Distributional fairness, Experiment |
| Date: | 2025 |
| URL: | https://d.repec.org/n?u=RePEc:hal:journl:hal-05165240 |
| By: | Frédéric Marty; Thierry Warin |
| Abstract: | This article examines the competitive implications of algorithmic pricing in digital markets. While algorithmic pricing can enhance market efficiency through real-time adjustments, personalized offers, and inventory optimization, it also raises substantial risks, including tacit collusion, discriminatory pricing, market segmentation, and exploitative consumer manipulation. Drawing on theoretical models, simulations, and emerging empirical evidence, the brief explores how algorithmic strategies may lead to supra-competitive prices without explicit coordination, particularly in oligopolistic or data-rich environments. It also highlights how common algorithm providers, shared data sources, and learning dynamics can undermine competition. Special attention is given to the challenges posed by loyalty penalties, ecosystem lock-in, and granular predatory pricing. The paper concludes with a set of policy recommendations emphasizing updated enforcement tools, transparency mechanisms, ex ante regulation for dominant platforms, and a coordinated approach to digital market oversight that balances innovation with consumer protection. Cet article examine les implications concurrentielles de la tarification algorithmique sur les marchés numériques. Si la tarification algorithmique peut améliorer l'efficacité du marché grâce à des ajustements en temps réel, des offres personnalisées et une optimisation des stocks, elle présente également des risques importants, notamment la collusion tacite, la tarification discriminatoire, la segmentation du marché et la manipulation abusive des consommateurs. S'appuyant sur des modèles théoriques, des simulations et des données empiriques émergentes, cet article explore comment les stratégies algorithmiques peuvent conduire à des prix supraconcurrentiels sans coordination explicite, en particulier dans les environnements oligopolistiques ou riches en données. Il souligne également comment les fournisseurs d'algorithmes communs, les sources de données partagées et la dynamique d'apprentissage peuvent nuire à la concurrence. Une attention particulière est accordée aux défis posés par les pénalités de fidélité, le verrouillage de l'écosystème et les prix prédateurs granulaires. L'article conclut par un ensemble de recommandations politiques mettant l'accent sur la mise à jour des outils d'application, les mécanismes de transparence, la réglementation ex ante des plateformes dominantes et une approche coordonnée de la surveillance du marché numérique qui concilie innovation et protection des consommateurs. |
| JEL: | L41 D43 L13 K21 G18 |
| Date: | 2025–08–04 |
| URL: | https://d.repec.org/n?u=RePEc:cir:circah:2025pr-09 |
| By: | Ajay K. Agrawal; Joshua S. Gans; Avi Goldfarb |
| Abstract: | 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 necessarily a complement to cognitive tools. Using these concepts, we can synthesise the empirical literature on the impact of computers and artificial intelligence (AI) on productivity and inequality. Specifically, while both computers and AI appear to increase productivity, computers have also contributed to increased inequality. Empirical work on the impact of AI on inequality has shown both increases and decreases, depending on the context. We also apply the model to understand how cognitive tools might influence incentives to automate processes and allocate decision-making authority within teams. |
| JEL: | D83 J24 L23 O33 |
| Date: | 2025–07 |
| URL: | https://d.repec.org/n?u=RePEc:nbr:nberwo:34034 |
| By: | Albanesi, Stefania; Da Silva, António Dias; Jimeno, Juan F.; Lamo, Ana; Wabitsch, Alena |
| Abstract: | We examine the link between the diffusion of artificial intelligence (AI) enabled technologies and changes in the female employment share in 16 European countries over the period 2011-2019. Using data for occupations at the 3-digit level, we find that on average female employment shares increased in occupations more exposed to AI. Countries with high initial female labor force participation and higher initial female relative education show a stronger positive association. While there exists heterogeneity across countries, almost all show a positive relation between changes in female employment shares within occupations and exposure to AI-enabled automation. JEL Classification: J23, O33 |
| Keywords: | artificial intelligence, employment, gender, occupations, skills |
| Date: | 2025–07 |
| URL: | https://d.repec.org/n?u=RePEc:ecb:ecbwps:20253077 |
| By: | Omar R. Malik |
| Abstract: | Using data from the U.S. Census Bureaus Business Trends and Outlook Survey (BTOS), I examine the adoption of AI among US firms at national, state, industry, and firm size levels. I find that adoption remains overall low (only around 7% of firms currently use AI), but is on a steady upward trajectory with a rising share of firms planning to implement AI. Adoption rates vary significantly across regions and sectors: some states are emerging as early adopters, while others lag, and knowledge-intensive industries (such as information technology and professional services) along with larger firms show higher openness to AI adoption compared to sectors like construction or small businesses. In general, these trends indicate that a quiet revolution in AI adoption is underway; a gradual but expanding diffusion of AI across the economy with important implications for future productivity and policy. |
| Date: | 2025–05 |
| URL: | https://d.repec.org/n?u=RePEc:arx:papers:2505.14721 |
| By: | Egana-delSol, Pablo (Adolfo Ibanez University); Vargas-Faulbaum, Luis (Adolfo Ibanez University) |
| Abstract: | This article offers a comprehensive review of Artificial Intelligence's (AI) effects on global labour markets, with a particular focus on developing economies. Drawing on an extensive body of evidence, it demonstrates that AI's disruptive potential diverges markedly from earlier waves of automation, extending its reach into occupations once deemed insulated—especially those demanding advanced education and complex cognitive abilities. The analysis reveals significant heterogeneity in AI exposure across countries at different development stages and among workers distinguished by skill sets, educational attainment, age, and gender, underscoring its unequal distributional consequences. To harness AI's benefits while safeguarding vulnerable groups, we propose four strategic policy levers: bolstering digital infrastructure, expanding vocational training and lifelong upskilling programmes, formalising labour markets, and integrating AI tools within social protection delivery. Collectively, these measures foster a human centred adoption of AI, bridge the digital divide, and promote inclusive growth, thereby mitigating adverse impacts on employment and wages. |
| Keywords: | artificial intelligence, labour market, inequality, automation, social protection |
| JEL: | J23 J24 J31 O1 O33 |
| Date: | 2025–07 |
| URL: | https://d.repec.org/n?u=RePEc:iza:izapps:pp216 |
| By: | Joshua Foster; Fredrik Odegaard |
| Abstract: | This paper proposes a new demand estimation method using attention-based language models. An encoder-only language model is trained in a two-stage process to analyze the natural language descriptions of used cars from a large US-based online auction marketplace. The approach enables semi-nonparametrically estimation for the demand primitives of a structural model representing the private valuations and market size for each vehicle listing. In the first stage, the language model is fine-tuned to encode the target auction outcomes using the natural language vehicle descriptions. In the second stage, the trained language model's encodings are projected into the parameter space of the structural model. The model's capability to conduct counterfactual analyses within the trained market space is validated using a subsample of withheld auction data, which includes a set of unique "zero shot" instances. |
| Date: | 2025–07 |
| URL: | https://d.repec.org/n?u=RePEc:arx:papers:2507.17564 |
| By: | Erdinc Akyildirim (University of Nottingham); Gamze Ozturk Danisman (Istanbul Bilgi University); Steven Ongena (University of Zurich - Department Finance; Swiss Finance Institute; KU Leuven; NTNU Business School; Centre for Economic Policy Research (CEPR)) |
| Abstract: | Using a panel of 929 U.S. publicly listed firms, this paper investigates the impact of artificial intelligence (AI) employment on the disclosure of political risk in corporate earnings calls. We utilize the firm-level AI employment measure developed by Babina et al. (2024), based on resume and job posting records. Furthermore, we supplement it with our newly generated AI disclosure indices at the firm level, created through textual analysis of earnings call transcripts. Our findings indicate that firms with greater AI employment are significantly less likely to disclose information about political risk during earnings calls. We propose a dual mechanism that underpins this association. First, AI enables narrative management: firms use AI tools to strategically alter the tone and wording of disclosures, avoiding phrases that may elicit unfavorable sentiment, leading to a reduction in reputational risk. Second, AI improves firms’ internal performance and risk management, hence reducing the need for voluntary political risk disclosures. Our findings add to the literature on voluntary disclosure and the economic implications of AI by indicating that AI, as a general-purpose technology, has unintended consequences for corporate transparency. |
| Keywords: | Artificial Intelligence (AI), political risk, voluntary disclosures, earnings calls, textual analysis, AI disclosure index |
| Date: | 2025–06 |
| URL: | https://d.repec.org/n?u=RePEc:chf:rpseri:rp2556 |
| By: | Giorgos Iacovides; Wuyang Zhou; Danilo Mandic |
| Abstract: | Opinions expressed in online finance-related textual data are having an increasingly profound impact on trading decisions and market movements. This trend highlights the vital role of sentiment analysis as a tool for quantifying the nature and strength of such opinions. With the rapid development of Generative AI (GenAI), supervised fine-tuned (SFT) large language models (LLMs) have become the de facto standard for financial sentiment analysis. However, the SFT paradigm can lead to memorization of the training data and often fails to generalize to unseen samples. This is a critical limitation in financial domains, where models must adapt to previously unobserved events and the nuanced, domain-specific language of finance. To this end, we introduce FinDPO, the first finance-specific LLM framework based on post-training human preference alignment via Direct Preference Optimization (DPO). The proposed FinDPO achieves state-of-the-art performance on standard sentiment classification benchmarks, outperforming existing supervised fine-tuned models by 11% on the average. Uniquely, the FinDPO framework enables the integration of a fine-tuned causal LLM into realistic portfolio strategies through a novel 'logit-to-score' conversion, which transforms discrete sentiment predictions into continuous, rankable sentiment scores (probabilities). In this way, simulations demonstrate that FinDPO is the first sentiment-based approach to maintain substantial positive returns of 67% annually and strong risk-adjusted performance, as indicated by a Sharpe ratio of 2.0, even under realistic transaction costs of 5 basis points (bps). |
| Date: | 2025–07 |
| URL: | https://d.repec.org/n?u=RePEc:arx:papers:2507.18417 |
| By: | Haochen Luo; Yuan Zhang; Chen Liu |
| Abstract: | Sparse portfolio optimization is a fundamental yet challenging problem in quantitative finance, since traditional approaches heavily relying on historical return statistics and static objectives can hardly adapt to dynamic market regimes. To address this issue, we propose Evolutionary Factor Search (EFS), a novel framework that leverages large language models (LLMs) to automate the generation and evolution of alpha factors for sparse portfolio construction. By reformulating the asset selection problem as a top-m ranking task guided by LLM-generated factors, EFS incorporates an evolutionary feedback loop to iteratively refine the factor pool based on performance. Extensive experiments on five Fama-French benchmark datasets and three real-market datasets (US50, HSI45 and CSI300) demonstrate that EFS significantly outperforms both statistical-based and optimization-based baselines, especially in larger asset universes and volatile conditions. Comprehensive ablation studies validate the importance of prompt composition, factor diversity, and LLM backend choice. Our results highlight the promise of language-guided evolution as a robust and interpretable paradigm for portfolio optimization under structural constraints. |
| Date: | 2025–07 |
| URL: | https://d.repec.org/n?u=RePEc:arx:papers:2507.17211 |
| By: | Benjamin Coriat; Eric Benhamou |
| Abstract: | This paper presents a novel hierarchical framework for portfolio optimization, integrating lightweight Large Language Models (LLMs) with Deep Reinforcement Learning (DRL) to combine sentiment signals from financial news with traditional market indicators. Our three-tier architecture employs base RL agents to process hybrid data, meta-agents to aggregate their decisions, and a super-agent to merge decisions based on market data and sentiment analysis. Evaluated on data from 2018 to 2024, after training on 2000-2017, the framework achieves a 26% annualized return and a Sharpe ratio of 1.2, outperforming equal-weighted and S&P 500 benchmarks. Key contributions include scalable cross-modal integration, a hierarchical RL structure for enhanced stability, and open-source reproducibility. |
| Date: | 2025–07 |
| URL: | https://d.repec.org/n?u=RePEc:arx:papers:2507.18560 |