Twenty five mathematicians who have each won the Fields Medal signed an open letter this week. The Fields Medal is widely seen as the top honor in mathematics.
The letter raises concerns about how AI companies are racing to solve famous math problems. The signatories say this competition is putting real strain on the research community.
Their warning comes after a public dispute involving OpenAI. NYU professor Tristan Buckmaster said the company pressured him not to credit a collaborator who works at Anthropic.
That collaborator helped solve a well known math problem. Buckmaster also questioned whether OpenAI used their shared work with its Codex tool to help produce its own proof.
OpenAI's Withdrawal From CalTech
OpenAI released a proof related to the Navier-Stokes equations after what it described as a weekend of intense computing work. The proof has not yet been verified by outside mathematicians.
On Thursday, OpenAI withdrew its sponsorship of a math event at CalTech. The move came after researchers at the university raised objections to the company's involvement.
OpenAI has not issued a detailed public explanation for the withdrawal. The company's math-related announcements have drawn scrutiny in recent weeks.
The mathematicians who signed the open letter say speed is becoming a problem. They wrote that solutions are often announced in a rush, without time for a proper writeup or a clear record of prior work.
They argue this creates real risks around attribution and plagiarism. Without mathematicians involved in reviewing and integrating new results, they say the findings cannot be fully understood or trusted.
Fear of a Race to Secrecy
Some mathematicians are now wondering if their own work, shared through tools like Codex, was fed back into training newer AI models. That uncertainty is fueling distrust across the field.
The concern is that labs with large budgets can now spend heavily on AI systems to reach a proof before human researchers do. That dynamic could push researchers toward keeping their work private instead of sharing it openly.
This is not the first document to address these issues. In June, a group of mathematicians released the Leiden Declaration, which looks at how AI proofs are changing the profession.
That document includes recommendations for mathematicians, universities, and policymakers as they adjust to AI's growing role in research.
The new letter draws a comparison to other fields facing similar pressure from AI tools, including software engineering. The signatories say the core question is not just who gets credit for a proof.
They argue the deeper issue is preserving the process that trains new mathematicians, generates new questions, and connects discoveries to the wider world. As of this week, OpenAI has not issued a new statement responding directly to the letter's authors.
