Large language models can write essays, summarize legal clauses, explain ancient history, draft emails, and produce code that looks impressively official. Then you ask one to multiply two awkward ...

The result is correct but challenges core norms of mathematics: checking proofs, crediting ideas and keeping research open to everyone.

In September 2024, OpenAI previewed a model that behaved differently from the AI systems most people had grown accustomed to.

Mathematician Will Sawin discusses his experience reviewing and refining a mathematical proof devised by OpenAI's internal model—and what that could mean for mathematics.