OpenAI’s feud with mathematicians is only escalating
TechCrunch Tim Fernholz ● Covered by 21 sources
Top mathematicians are accusing AI labs of racing to steal credit for major proofs. OpenAI just got dragged into it, and the fight is now about who gets to own math.
Based on reporting by TechCrunch, Tim Fernholz — read the original for the full story.
Summary, retelling and take written by AI under human oversight; images are AI-generated illustrations. How we work · Report an error
A group of 25 leading mathematicians has published an open letter saying AI labs are threatening their work as they try to outdo one another on famous math problems. Every signer has won the Fields Medal, which is about as close as mathematics gets to a gold standard for prestige.
The pressure has already turned personal. NYU professor Tristan Buckmaster said this week that OpenAI pushed him not to credit a collaborator at Anthropic for solving an important problem, and he also questioned whether the company used work tied to Codex to produce its own proof over a long weekend of inference. OpenAI’s proof is still unverified.
Then came another flashpoint: on Thursday, OpenAI pulled sponsorship from a math event at CalTech after researchers there criticized the company. That’s a neat little summary of the mood right now — distrust first, collaboration second, if at all.
The letter’s bigger complaint is not just about credit lines. The signers argue that rapid-fire announcements leave too little time for proper writeups, for teasing apart new methods, and for giving previous work its due. They also warn that if AI-conceived ideas don’t get taken up by willing mathematicians, those ideas never really enter the mathematical canon.
And the fear is spreading. Some researchers now worry that Codex use may have fed into OpenAI’s newer models, which makes the whole culture of open research feel less open and a lot more paranoid. The signers say frontier labs can spend tens of millions of dollars to use LLMs to beat the original researchers to a proof, and that incentive pushes everyone toward secrecy.
This letter follows June’s Leiden Declaration, which also tried to map out how LLM proofs could change mathematics and offered recommendations for mathematicians, institutions, and policymakers. The deeper point is simple: in math, the proof matters, but so does the machinery around it — the students, the questions, the methods, the chain of human transmission that keeps the field alive.
My take — AI-written commentary, not fact-checked reporting
This is what happens when AI companies treat research like a scoreboard and act surprised when the research community starts keeping receipts. The real damage isn’t just attribution drama; it’s the chill factor, because secrecy is poison in fields that depend on shared methods and trust. OpenAI didn’t invent that problem, but it’s helping make it louder and uglier.
Read more about this at: TechCrunch
