MS&E 306: Seminar on Economic Foundations of AI Alignment

3 units · Letter or Credit/No Credit

Explores recent research on AI alignment from an economic perspective, motivated by the prospect that increasingly capable AI systems may escape meaningful human control. Draws on social choice, principal-agent theory, mechanism design, and related foundations. Possible topics include preference aggregation; strategic agent behavior; oversight, auditing, and control; and liability and insurance as tools for mitigating catastrophic risk. Doubles as an exploration of how we learn with AI. Students lead weekly discussions of papers, presenting both what they learned and how they learned it, and complete a quarter-long research project. Examines a range of techniques for using AI in the learning process, with the goal of developing and sharing best practices. Enrollment is by application only and limited to 15 students, with preference to Ph.D. students. Graduate background in at least one of microeconomic theory, game theory, optimization, or machine learning is expected.

Offered in Winter 2027 at Stanford University.

Winter 2027 sections

More MS&E courses

All MS&E courses · All departments