MLESI Recap 2025

The 2025 Machine Learning in Economics Summer Institute (MLESI), hosted by the Center for Applied Artificial Intelligence (CAAI) at the University of Chicago Booth School of Business, brought together faculty, researchers, and students for an intensive week of discussion, collaboration, and discovery at the intersection of economics and machine learning. Taking place on the University of Chicago campus, the Institute provided participants with an opportunity to engage with cutting-edge research, learn from leading scholars in the field, and share their own work with peers.
The program opened with a welcome dinner for students, setting the stage for a week of connection and discussion. The first day began with introductions and an overview session before diving into foundational topics in supervised learning, prediction policy problems, and the theory of prediction. In the afternoon, discussions turned to emerging developments in large language models and their applications in economics. Each day ended with office hours where students had the opportunity to connect directly with faculty mentors.

Over the following days, the Institute showcased a diverse set of perspectives on how machine learning tools can shape and inform economic research. Sessions featured topics such as  machine learning for treatment heterogeneity (Stefan Wager), structural approaches to economic modeling (Sanjog Misra), what artificial intelligence “understands” (Keyon Vafa) and the role of cognitive science in designing human-AI collaborations (Ted Sumers). 

MLESI also placed strong emphasis on community. Participants took part in a Chicago River architecture boat tour, group dinners at local, and casual networking opportunities throughout the week.
The conference component of the week (MLESC) featured a mix of longer research talks and flash presentations. Topics ranged widely, from demand estimation using text and image data (Giovanni Compiani) to copyright and competition in markets shaped by unstructured data (Sukjin Han). Flash talks highlighted the creativity and breadth of conference research. These sessions underscored the wide range of work in the field of machine.

The keynote addresses were among the highlights of the week. Nikhil Agarwal delivered a lecture on sufficient statistics, while Jon Kolstad presented on cognitive capacity and medical diagnosis, framing broader questions about the role of AI in decision-making processes. Both keynotes provoked thoughtful discussion about the evolving relationship between economics, data, and artificial intelligence.

The final day of the Institute brought together the many threads of the week’s discussions. Ralph Koijen presented on representations of firms and investors, while Alex Imas led a session on the social science of AI. A concluding panel discussion provided space for participants to reflect on the themes of the week.

The 2025 Machine Learning in Economics Summer Institute was a resounding success, blending rigorous academic content with opportunities for professional development and community building. Through a carefully designed mix of lectures, workshops, keynotes, and informal gatherings, the Institute not only highlighted groundbreaking research but also fostered connections among scholars who will continue to shape the future of economics and artificial intelligence. The event demonstrated the power of bringing together diverse perspectives and experiences, reaffirming MLESI’s role as a vital hub for advancing this evolving field.