OpenAI posts 722 AI-generated math preprints, pulls three

OpenAI uploaded 722 AI-generated math preprints to a GitHub repository on October 6, Nature reports, covering purported progress on 372 open problems in geometry, algebra and computer science. The papers were produced by an unreleased internal model pointed at roughly 4,000 problems. Only 162 of the 722 manuscripts had their main results formalized in Lean; the rest went out without independent verification.

A day later, OpenAI pulled three preprints after a sign error invalidated one argument and the constructions in two dependent papers, and revised 14 other manuscripts with proof repairs and corrected statements, Retraction Watch reported.

The response from working mathematicians was unusually blunt. Alvaro Lozano-Robledo, a number theorist at the University of Connecticut, called the drop 'staggering' and asked on X, "Why is OpenAI trying to solve hundreds of math problems to be released all at once?" Ursula Martin of Oxford University described the approach as "tossing the community a messy first draft and expecting us to clean it up for them." Michael Harris, a number theorist at Columbia University, said flatly: "I can't bring myself to spend my time studying [these preprints], so I will not be reading these papers." A group letter circulating on social media says that "Releasing over 700 files at once is not a demonstration of scholarship, but a demonstration of power."

Not every mathematician objected. Levent Alpöge, a mathematician at Anthropic, acknowledged that researchers whose projects might be "pre-empted by the OpenAI release" would have "sad stories," yet called the moment "the most significant moment in mathematical history." OpenAI said its goal is to "provide tools for mathematicians and scientists" and that it evaluates models "against difficult frontier problems in all sciences." The batch did not include solutions to the five remaining Millennium Prize Problems, though the company had claimed progress on a Navier-Stokes problem in September. Two of the researchers we track in our Who's Who directory shared the Nature piece.