A group of 25 prominent mathematicians, all recipients of the prestigious Fields Medal, have publicly expressed concerns about how AI research labs, including OpenAI, are handling breakthroughs in solving complex mathematical problems. They argue that the rush to publish AI-generated proofs risks undermining the collaborative and transparent nature of mathematical research.
Recently, NYU professor Tristan Buckmaster accused OpenAI of pressuring him not to credit a collaborator from Anthropic who contributed to solving a significant math problem. Buckmaster also questioned whether OpenAI used insights from their work with Codex to develop its own proof during an intensive weekend of AI inference.
In response to criticism from CalTech researchers, OpenAI withdrew its sponsorship of a mathematics event at the university. The mathematicians emphasize that while AI's ability to tackle major mathematical challenges could benefit humanity, these solutions must be comprehensible and properly integrated into the broader mathematical community.
The letter highlights concerns about the speed at which AI-generated proofs are announced, often without thorough documentation, isolation of new methods, or appropriate citation of prior work. This raises questions about attribution and potential plagiarism. The mathematicians stress that human involvement remains essential for validating, understanding, and transmitting new ideas within the discipline.
There is growing unease among mathematicians about whether their use of AI tools like Codex is being incorporated into newer AI models without consent, potentially threatening the open research culture. With significant funding available, AI labs might prioritize secrecy to secure discoveries before traditional researchers can publish, altering the dynamics of academic collaboration.
This letter follows the Leiden Declaration issued earlier this year, which also addressed the impact of large language models on mathematical research and offered guidance for the community and policymakers.
Mathematicians argue that the value of their work extends beyond individual proofs to the intellectual framework that fosters education, innovation, and cultural integration. The challenges faced by mathematicians reflect broader issues confronting scientific and creative fields as AI reshapes workflows.
The community warns that as AI transforms how work is done, it is crucial to maintain focus on the fundamental goals of research and creativity to ensure progress benefits society as a whole.