Maximilian Kasy’s Conference – The Means of Prediction, how AI Really Works (and Who Benefits)
Dear all,
We are delighted to invite you to a conference with Maximilian Kasy, renowned economist at the University of Oxford. Maximilian Kasy will first engage at CREST during the Applied micro seminar and later present his new book, The Means of Prediction, how AI Really Works (and Who Benefits).
This special event is organized by CREST, by Julien Combe and Étienne Ollion and will include a cocktail reception after the conference to foster discussions and networking.
? Date: May 5, 2026
? Time: 6PM – 8PM
? Location: ENSAE Paris – Amphi 200 (the cocktail reception will be held in the great hall)
? Registration: https://forms.gle/NFSx99UHed6Kybp76
About the Conference:
In The Means of Prediction, Maximilian Kasy offers a clear and critical introduction to how artificial intelligence and machine learning actually work. Rather than focusing on technical complexity, the book explains AI as a system of automated prediction shaped by data, objectives, and human choices. Kasy introduces the notion of the “means of prediction” — the data, computing power, and institutions that determine who controls and benefits from AI. He argues that the central issue is not machines replacing humans, but who sets the goals that AI systems pursue. The book makes a strong case for democratic oversight of AI to ensure its benefits are shared broadly across society.
Provisional program:
- Maximilian Kasy’s presentation
- Discussion about The Means of Prediction, with Étienne Ollion
- Open discussion with the audience
About Maximilian Kasy:
Maximilian Kasy holds a PhD in Economics from UC Berkeley and is Professor of Economics at the University of Oxford and coordinator of the Machine Learning and Economics Group. His research spans machine learning theory and statistical decision methods, with a strong focus on the social and political economy of algorithmic decision-making. He works on issues of economic inequality, basic income and job-guarantee programs, and causal inference, as well as on publication bias, experimental design, and the role of statistics in scientific practice.
We look forward to seeing you.