Key facts
- Twenty-five Fields Medal winners signed an open letter concerning AI's impact on mathematics.
- The mathematicians argue AI labs rush solutions, raising attribution and plagiarism concerns.
- NYU professor Tristan Buckmaster accused OpenAI of pressuring him regarding collaborator credit.
- OpenAI withdrew sponsorship from a CalTech math event.
- The signatories believe AI-conceived ideas need human integration to become part of the mathematical canon.
A group of 25 mathematicians, each a recipient of the prestigious Fields Medal, have signed an open letter expressing concern that artificial intelligence labs are jeopardizing the integrity of mathematical research. The signatories argue that AI companies, in their race to solve complex mathematical problems, are rushing solutions without proper write-ups, hindering the isolation of new methods and the citation of previous work, which raises significant attribution and plagiarism issues.
One specific incident involved NYU professor Tristan Buckmaster, who alleged that OpenAI pressured him not to credit a collaborator from Anthropic for solving a significant math problem. Buckmaster also questioned whether OpenAI utilized his and his collaborator's work with Codex to produce its own proof over a weekend.
In response to criticism from researchers at CalTech, OpenAI withdrew its sponsorship of a math event hosted by the university. The mathematicians behind the letter emphasize that while AI-generated solutions could benefit humanity, they must be understandable and communicable to the broader mathematical community. They fear that the current dynamic, where labs can spend heavily on LLMs to beat original researchers to proofs, will incentivize secrecy and threaten the culture of open research.
This development follows the June release of the Leiden Declaration, which also addressed the impact of LLM proofs on mathematical work and provided recommendations for mathematicians, institutions, and policymakers. The mathematicians draw parallels between their situation and other creative and scientific fields, warning that the issues they face could extend to all of humanity as AI transforms workflows.