Harvard Math Professor Joins Independent Group Advising AI Companies on Research Standards

Harvard Mathematics professor Melanie Matchett Wood has joined an independent group of nine mathematicians that will advise OpenAI and other AI companies on how to responsibly release mathematical results, the group announced Monday.

The Advisory Group on Mathematics and Artificial Intelligence is hosted at the Institute for Advanced Study, an independent research center in Princeton, New Jersey. Wood is its only Harvard affiliate, and its members are unpaid and plan to publish their recommendations publicly.

Wood said in a Monday interview that the group would offer advice to any AI company that wants it.

“Our goal is to work for the best interests of mathematics and the future of mathematics, and to do that by advising any AI companies that need such advice on their interactions with math research,” she said.

The announcement came the same day OpenAI said an internal model had resolved the Navier-Stokes Millennium Prize problem and more than 100 long-standing open problems, a pace the company wrote had surprised its own mathematicians. The group’s immediate task is advising OpenAI on how to coordinate the release of those results.

It also follows a Sept. 11 declaration signed by 25 Fields Medalists, which argued that using famous open problems as AI benchmarks damages mathematics.

Wood said AI output falls short of basic scholarly norms.

“AI systems are able to give a lot of output that is mathematical in some ways, but is not conforming to important academic standards,” she said. “It often includes plagiarism. It fails to present ideas clearly, fails to document the entire scientific process, and usually there’s no human who is willing to take responsibility for the output.”

The field has no established norms for handling such output, Wood said, even though some of it could move the study of mathematics forward.

The advisory group published a feedback form on its website Monday and has already heard from mathematicians, Wood said, though she argued it was too early to say what the new norms should be.

“There are going to be some hard answers that we’re going to have to work out,” she said.

She added that the group formed after OpenAI approached several mathematicians about an advisory board of its own. They declined and organized independently.

“An independent group would be much better positioned to work for the needs of the mathematical community,” she said. “We didn’t want to have any special relationship with OpenAI. Whatever work we do and advice we give, we want to be able to give to any company.”

Wood declined to weigh in on broader guardrails for the industry — days after Anthropic CEO Dario Amodei called on AI companies to slow the pace of model development — saying she was “not expert enough” to speak to AI systems in general.

But she said her field offered a preview of the wider debate.

“What we see in mathematics — that there are both opportunities and challenges and risks — is a microcosm for what one sees in society,” Wood said.

—Staff writer Francesco Palma can be reached at [email protected].