Twenty-five top mathematicians have penned an open letter criticizing AI labs, particularly OpenAI, for jeopardizing the integrity and collaborative culture of mathematical research as AI tools increasingly target famous unresolved problems.
- Fields Medalists claim AI threatens proper credit and collaboration in math.
- OpenAI accused of pressuring researchers over proof attribution.
- Concerns grow about secrecy and verification in AI-generated mathematical work.
What happened
Twenty-five distinguished mathematicians, all recipients of the Fields Medal, have authored an open letter expressing alarm that AI labs, notably OpenAI, are compromising the collaborative and transparent nature of mathematical research. These concerns intensified after Tristan Buckmaster, a professor at NYU, accused OpenAI of pressuring him not to acknowledge a collaborator from rival AI company Anthropic in a significant recent mathematical proof. Furthermore, OpenAI abruptly ended its sponsorship of a mathematics event at CalTech in response to criticism from university researchers.
The mathematicians argue that, while AI’s ability to solve challenging mathematical problems could revolutionize the field, the current rush to announce results often lacks thorough writeups, isolation of new ideas, and appropriate citations. This situation raises serious issues about attribution and potential plagiarism. Without mathematicians actively documenting and integrating AI-generated ideas, the essential human transmission and understanding of mathematical discoveries could be lost.
Why it matters
The debate underscores a critical tension between rapid AI-driven discoveries and the traditional processes of mathematical validation, communication, and credit assignment. Mathematics relies heavily on the intellectual framework built around proofs, which includes mentorship, research culture, and integration into the broader scientific canon. If AI labs prioritize speed and secrecy, it risks undermining these foundations and could discourage open collaboration essential for progress.
Moreover, mathematicians fear that AI models may be trained on their unpublished or collaborative work without consent, breeding mistrust. This threatens the culture of openness that has long been vital to scientific advancement. The escalating conflict hints at broader challenges posed by AI across all scientific and creative fields—balancing innovation speed with transparency, fairness, and human stewardship.
What to watch next
The mathematical community’s response to AI will continue evolving as institutions, policymakers, and researchers seek standards that ensure AI contributions are both verifiable and properly credited. The earlier Leiden Declaration already offered guidelines addressing AI’s role in math, and the current letter adds urgency to developing practical frameworks that protect the community’s norms.
Observers should monitor how AI labs respond to the accusations and whether they adopt more transparent practices. Additionally, keep an eye on emerging policies that govern AI use in research and the broader scientific ecosystem. This debate foreshadows how other disciplines might handle rapid AI-driven discoveries and preserve their core values.