gecon.ai

Economics of Generative AI Research Lab

Abstract

Economics of Generative AI Research Lab (gecon.ai) is an independent nonprofit research organization dedicated to studying artificial intelligence, economics, and related social questions. Our work focuses on public-interest research, open publication, and rigorous methods for understanding AI systems as both technologies and objects of study.

Goals

Economics of Generative AI Research Lab pursues three primary goals that guide our research agenda and shape our contributions to the field. These goals reflect our commitment to understanding the complex relationship between artificial intelligence and economic systems, both in theory and practice.

Exploring Effects of Generative AI on Economics. We investigate how generative AI technologies transform economic theory, policy, and practice. This includes examining how AI-driven tools reshape market dynamics, alter economic modeling approaches, and create new opportunities for understanding complex economic phenomena. Our research explores both the positive impacts, such as enhanced predictive capabilities and more sophisticated economic simulations, as well as potential challenges including algorithmic bias, market manipulation, and the need for new regulatory frameworks.

Effects of AI on Human Behaviour. We study how artificial intelligence systems influence human decision-making, preferences, and behavioral patterns in economic contexts. This research examines whether exposure to AI recommendations changes how individuals make economic choices, how AI-mediated interactions affect trust and cooperation in economic settings, and how the presence of AI systems might alter traditional economic behaviors. Understanding these effects is crucial for predicting the long-term impact of AI integration into economic systems.

Exploring Ways to Evaluate LLMs as Subjects of Behavioural Research. We develop methodological frameworks for treating Large Language Models (LLMs) as research subjects in behavioral economics experiments. This involves creating standardized protocols for testing economic preferences, decision-making patterns, and behavioral responses in LLMs. Our goal is to establish rigorous evaluation methods that allow researchers to systematically study AI systems' economic behaviors, compare them across different models, and understand how these behaviors emerge from the underlying architectures and training processes.

Mission

The mission of Economics of Generative AI Research Lab (gecon.ai) is to conduct independent scientific and educational research on artificial intelligence, technology, economics, and related social questions for the public benefit. Our current emphasis is on large language models, AI agents, and the emerging agent economy.

We publish papers, datasets, software, benchmarks, evaluations, and educational materials intended to strengthen public understanding and improve research quality in a rapidly developing field.

Although AI is our primary focus, we also support other board-approved public-interest research projects that advance educational and scientific purposes.

Research Projects

Large Language Models exhibit consistent economic preference structures that can be systematically evaluated through experimental economics methodologies.

We investigate whether LLMs demonstrate consistent preference structures, risk attitudes, and decision-making patterns that mirror or diverge from human economic behavior across various economic choice scenarios.See full proof.

Large Language Models exhibit values and behaviors characteristic of WEIRD (Western, Educated, Industrialized, Rich, and Democratic) populations in economic decision-making contexts.

We examine the cultural and ethical biases embedded in LLM responses and decision-making processes, evaluating the extent to which these systems reflect the cultural biases of their training data in economic contexts.See full proof.

AI-generated media is indistinguishable from human-generated media under controlled experimental conditions.

Through blind evaluation protocols, we investigate whether human evaluators can reliably distinguish between AI-generated and human-generated media, examining the implications for media authenticity and the validity of using AI systems in content generation tasks.See full proof.

✦ see all projects ✦

Collaborators

Nikita Tkachenko Nikita Tkachenko is a collaborator at Economics of Generative AI Research Lab. His work focuses on artificial intelligence, economics, and the emerging agent economy, with particular interest in research design, evaluations, and public-interest scientific output.

Jake Cosgrove Jake Cosgrove is a collaborator at Economics of Generative AI Research Lab and serves as Secretary and Treasurer of the corporation. He supports the organization’s governance, operations, and financial oversight.

Parsa Rahimi Parsa Rahimi is a collaborator at Economics of Generative AI Research Lab and serves on the board of directors. His involvement supports the organization’s governance and public-interest research mission.

Alessandra Cassar Alessandra Cassar is an affiliated scholar at Economics of Generative AI Research Lab and Professor of Economics at the University of San Francisco. Her research uses laboratory and field experiments to study human behavior, with interests spanning evolutionary processes, female competitiveness, conflict and disaster victimization, prosociality, and social networks. She is also a research affiliate at ESI at Chapman University, a faculty affiliate at CEGA at UC Berkeley, vice-president of North America ESA, and an associate editor for the European Economic Review.

Jesse Anttila-Hughes Jesse Anttila-Hughes is an affiliated scholar at Economics of Generative AI Research Lab and Associate Professor and Chair of Economics at the University of San Francisco. His work combines physical and statistical approaches to economic, health, and social questions, with research on coupled human-environmental systems, climate impacts on society, planetary public health, and physical approaches to economics. He has also been recognized at USF for graduate student mentorship and for teaching innovation involving AI and LLM methods in econometrics.

✦ see all collaborators ✦

Collaborate

We welcome collaborations with researchers, institutions, and organizations interested in the intersection of artificial intelligence and economics. Whether you are working on related research, have data or resources to share, or are interested in joint projects, we would be delighted to hear from you. ⊕ We welcome collaboration opportunities even outside the generative AI space.

Our lab is particularly interested in collaborations involving behavioral economics, experimental economics, computational economics, and AI ethics. We are open to partnerships that can advance our understanding of how AI systems can be used as both tools and subjects in economic research.

If you would like to collaborate with us, please contact us at natkachenko@usfca.edu.