AI-Assisted Riemann Proof Claim Remains Unverified
Anthropic mathematician Levent Alpöge says he worked with Claude for six weeks to prove the Riemann hypothesis, a landmark conjecture about the distribution of prime numbers and a Clay Mathematics Institute Millennium Prize Problem. He released a short explanation and a large proof encoded in Lean; researchers report that its modules compile and it has passed several computer-based verification checks. These checks do not amount to broad mathematical acceptance, and experts say independent scrutiny and understanding of the proof will take time. Separately, Stanford researcher Alessandro Masullo has proposed a human-developed proof of the generalized Riemann hypothesis, which has drawn questions from reviewers. If Alpöge’s claim is confirmed, it would be a major mathematical achievement and a notable test of AI’s role in research.





