AI and Mathematics

How should we learn, prove, and do mathematics now?

Xiang Huang
Department of Computer Science, University of Illinois Springfield, 2026

• View or download the talk slides
• Beamer LaTeX source package (zip) · ai-and-mathematics-talk.tex

AI is changing how mathematics is done. In my recent work, I have used large language models to formalize mathematical proofs in Lean, a theorem prover and programming language. These proofs come from textbooks, research papers, and my own research. Meanwhile, AI has largely autonomously formalized Fermat’s Last Theorem in Lean, and AI-assisted work has tackled the Navier–Stokes Millennium Problem. As AI becomes so good at mathematics, how should we educate students, and how should we learn, prove, and do mathematics ourselves?