How AI Could Change the Way Mathematics Is Done

Summary

This essay examines what artificial intelligence could mean for the way mathematics is practiced. It begins with OpenAI’s September 8 claim that it had solved the Navier-Stokes problem after deploying 10,000 AI agents for 88 hours. The article says two teams settled the long-standing question by proving that the Navier-Stokes equation can exhibit singularities, with one team made up of AI agents and the other consisting of human mathematicians using AI technology. Rather than focusing on the technical details, the author asks whether AI assistance will change how mathematicians are educated and what counts as mathematical work. Because a mathematical result affects later results, the author argues that AI-generated or AI-assisted proofs cannot be isolated from human mathematics in the way computer chess games can be separated from human games. The essay suggests that AI may become better than humans at developing proofs, while freeing mathematicians to explore questions, conjectures, explanations, and broader structures. It also argues that modern mathematics has heavily rewarded rigorous, detail-oriented proof work, sometimes at the expense of speculation, intuition, play, and aesthetic insight. The author presents imaginative mathematicians such as René Thom as a possible model for a future in which humans contribute ideas and vision while AI helps establish formal validity. The conclusion remains speculative: machines may eventually address profound questions, but choosing meaningful questions remains a human responsibility for now.