Twenty-five Fields Medal-winning mathematicians have signed an open letter warning that the race among AI labs to solve major mathematical problems could weaken open research, attribution and intellectual credit.
The letter, titled “A Severe Misalignment of AI in Mathematics,” argues that AI could become a powerful tool for the field, but only if its results can be understood, explained, independently verified and connected to existing work.
The researchers said mathematics is about more than producing a correct proof. Progress also depends on understanding methods, recognising contributions, sharing ideas and passing knowledge to future researchers. They warned that rapid AI-generated breakthroughs may leave too little time for documentation, citation and proper review.
The concerns come amid a dispute involving OpenAI and mathematicians working on problems related to the Navier–Stokes equations. NYU professor Tristan Buckmaster accused the company of pressuring him over credit for collaborator Levent Alpöge. OpenAI has denied that Buckmaster’s earlier Codex interactions influenced its own work.
OpenAI also recently announced that an internal AI system produced a solution related to the Navier–Stokes problem, one of the Millennium Prize Problems. The proof has not been independently verified.
The signatories said concerns over AI labs reaching discoveries before researchers could encourage secrecy and weaken collaboration. They added that the debate could serve as an early warning for other knowledge-based professions.
Source: EntArabi
