OpenAI's Navier-Stokes Claim Triggers Math Credit Backlash Over 1 Prize Problem

1 articles · Updated · The Guardian · Sep 20

8 September's claim that OpenAI agents solved the Navier-Stokes problem has set off what mathematicians describe as an existential crisis over who deserves credit for AI-generated breakthroughs.
Critics say the result rests on human mathematical work absorbed by large language models, and that OpenAI's paper cited sources without adequately crediting researchers seen as close to a solution.
Tristan Buckmaster said OpenAI may also have benefited from work he pursued on Navier-Stokes using Codex; OpenAI denied directly accessing it but said it could not rule out product-use data helping improve the model.
The dispute has sharpened a broader complaint that AI firms use mathematics as a showcase while still depending on mathematicians to verify, interpret and apply often opaque outputs.
Many mathematicians still see AI as genuinely powerful and potentially useful, but the episode has intensified calls to separate the technology from the companies controlling it and to democratize access.