OpenAI’s math solutions aren’t meeting the field’s standards yet

What happened: OpenAI released hundreds of claimed solutions to hard math problems, saying it consulted an advisory group of mathematicians. The Advisory Group on Mathematics and Artificial Intelligence said it is up to the mathematical community to assess whether its recommendations were followed. Key numbers: 719 manuscripts released, 10 included Why it matters: A new paper documents discrepancies between natural language proofs and Lean code for an OpenAI solution derived from the Navier-Stokes equations. Mathematicians say autoformalized proofs should not be trusted without peer review. Coverage: 4 outlets in 24h

OpenAI released hundreds of claimed solutions to hard math problems, saying it consulted an advisory group of mathematicians. The Advisory Group on Mathematics and Artificial Intelligence said it is up to the mathematical community to assess whether its recommendations were followed.

A new paper documents discrepancies between natural language proofs and Lean code for an OpenAI solution derived from the Navier-Stokes equations. Mathematicians say autoformalized proofs should not be trusted without peer review.