Fields Medalists Warn AI Labs Are Undermining Mathematics Culture
Twenty-five leading mathematicians, all winners of the prestigious Fields Medal, sign an open letter warning that AI labs threaten their intellectual work as they race to solve famous math problems. The letter lands amid an escalating feud between OpenAI and the mathematics community. This week, NYU professor Tristan Buckmaster accuses OpenAI of pressuring him not to credit a collaborator who works for Anthropic, and questions whether the company used their Codex work to produce its own groundbreaking proof over a marathon weekend of inference. OpenAI also withdraws its sponsorship of a math event at CalTech after criticism from researchers there.
The signatories argue that AI solutions only benefit humanity if the math community can understand and communicate them. Instead, proofs are often announced in a rush, leaving no time for proper writeups, the isolation of new methods, or citation of previous work. OpenAI's own proof remains unverified. The letter raises severe attribution and plagiarism concerns, noting that AI-conceived ideas require willing mathematicians to develop and integrate them into the mathematical canon, or the crucial human transmission chain between mathematicians will be lost.
Fear is spreading among researchers, with some wondering whether their Codex use is being fed into OpenAI's new models. Because frontier labs can spend tens of millions of dollars using LLMs to beat original researchers to a proof, the dynamic incentivizes secrecy and threatens the culture of open research. The letter follows the Leiden Declaration, released in June, which offers recommendations for mathematicians, institutions, and policymakers. As in other fields being transformed by AI, mathematicians point out that the value of math lies not just in proofs and credit, but in the intellectual work around them.