After turning a string of spectacular mathematical breakthroughs into a major reputational crisis, OpenAI is consulting human mathematicians to help it figure out a less disastrous path forward. The artificial intelligence giant’s aggressive entrance into formal mathematical research—highlighted by high-profile claims, including the resolution of the Navier-Stokes Millennium Prize problem—has frequently clashed with the traditions, culture, and professional norms of the global mathematics community.
On Monday, the company announced the formation of a new independent panel of mathematicians tasked with advising it and other AI companies on their interactions with mathematical research and the wider mathematics community. This oversight is intended to cover how new results are presented and released to the public and academic spheres. However, the abrupt arrival of the initiative caught many mathematicians by surprise. While researchers told The Verge that the group is a good first step, many said they were left with basic questions about what the panel will actually do, how much influence it will wield, and whether OpenAI will genuinely listen to its guidance. Further concerns were raised over whether such a small group of prominent researchers can adequately represent the wider mathematical community.
The nine-member group is stacked with individuals whom researchers described as celebrities in the field. Drawn from prestigious institutions including Stanford, Harvard, Oxford, and Cambridge, the roster boasts multiple Fields Medals and MacArthur “genius grants” among its members. The panel will be hosted by Princeton, New Jersey’s Institute for Advanced Study (IAS), one of the world’s most revered centers of mathematical research and the former intellectual home of towering historical figures including Albert Einstein, John von Neumann, and J. Robert Oppenheimer.
OpenAI stated that the group, formally called the Advisory Group on Mathematics and Artificial Intelligence (AGMAI), will operate independently. According to the company, the panel will have the freedom to offer advice that OpenAI has not explicitly requested, publicly comment on the impact of artificial intelligence on mathematics, and make its guidance public, although specific details on what this might entail remain undefined. It appears that the group will be granted early access to OpenAI’s internal research in order to fulfill its mandate. Members will not be paid by OpenAI, and the group retains the autonomy to change its membership as it sees fit.
"Its value depends on its members being able to exercise their own judgement and challenge ours," OpenAI stated, outlining that the group will advise on the review and communication of emerging results, help assess their significance, coordinate dissemination, and uphold academic and professional standards of mathematical research. Additionally, the panel is expected to advise on how AI tools can support mathematical research and learning.
Despite OpenAI’s emphasis on independence, it remains unclear how much real-world influence the group will actually wield or why it was formed in the first place. Simon Machado, a researcher at ETH Zurich in Switzerland, remarked that the entire announcement seemed "shrouded in mystery" and questioned how the nine specific members were selected. The group’s bare-bones website notes that it "came together after OpenAI approached some of its members about establishing an external advisory board," but that, following discussions with the company, the mathematicians decided to launch an independent group instead and invite others to join. It remains unclear which members were initially approached by OpenAI or how the remaining participants were ultimately chosen.
In a blog post published Tuesday, advisory group member Martin Hairer elaborated on the group’s origins and the thinking behind it, responding to what he termed a "torrent of misinformation" surrounding its formation. Hairer wrote that it is obvious AI has already had a profound impact on mathematical research, raising numerous complex questions regarding the correct attribution of ideas, priority, and human understanding of mathematical concepts. He added that there is no sense in the academic community taking a hard "ostrich" approach of simply ignoring AI labs and refusing to communicate with them on principle.
While acknowledging that the group was formed in response to an initial approach from OpenAI, Hairer stressed that not all of its members had been contacted by the company and that the resulting organization is genuinely independent of OpenAI and any other frontier AI labs. Beyond obvious confidentiality requirements, he noted that members have not signed any restrictive agreements governing what they can say publicly, that they receive no financial compensation, and that they are relying purely on the IAS for technical support.
Hairer, a professor of mathematics at Imperial College London and the Swiss Federal Institute of Technology in Lausanne (EPFL), emphasized that the overarching aim is to represent the interests of the mathematical community as well as humanly possible. While noting that the group is hardly a representative statistical sample of the broader field, he stated he is absolutely certain that all nine members take the responsibility extremely seriously and are acting in good faith.
Regardless of how the members were selected, the group currently comprises an elite cadre of academic mathematicians. This small size raises further questions about whether the panel possesses the necessary breadth of expertise required to evaluate everything coming its way, given the highly specialized skills and deep knowledge required to operate at the absolute frontiers of modern mathematics.
Francesco Fournier-Facio, an incoming mathematics professor at Heriot-Watt University in Scotland, described himself as fairly ambivalent about the panel, stating that greater engagement with mathematicians can only make things better. However, he questioned whether the group fully reflects the community whose work and livelihoods are being upended by OpenAI’s ongoing quest to rack up mathematical trophies. It feels like an ivory tower, he told The Verge.
Machado similarly questioned whether the group’s members fully understand, or can effectively speak to, the concerns of the broader mathematical community when it comes to the rapid rise of AI. Remarking that while they are amazing mathematicians, Machado expressed uncertainty over whether they are the ideal individuals to represent general interests in more political matters. For mathematicians in elite positions, he noted, the reality they face is very different from the reality confronting most working mathematicians today.
Whether the group can successfully earn the trust of the wider mathematical community will ultimately determine how effective it will be at containing the fallout from OpenAI’s recent parade of missteps in announcing results, and more importantly, softening the blow for whatever the company is planning to unveil next. In his blog post, Hairer admitted that members would be very naive to believe that AI labs will not try to spin whatever the panel says in a way that suits their public relations machine, which easily dwarfs anything the academic advisory group could possibly mount.
Nevertheless, Hairer argued it would be hypocritical not to engage now that OpenAI appears to be making a genuine effort, particularly after so many scholars criticized the company’s past conduct and poor communications. The group has yet to settle on the exact advice it intends to provide, he said, but it is responding to a clear and present problem. It is a fact, he noted, that AI companies have in recent months been producing high-profile mathematical results, and their publication and dissemination have fallen far short of acceptable mathematical practice, whichever way that standard is defined.
For now, Hairer stated that the group is actively gathering input through conversations with academic colleagues, an open public feedback form, and ongoing discussions on Proofs and Prompts, a blog that has quickly emerged as a key forum for mathematicians reckoning with the disruptive impact of artificial intelligence on their field.
Yet the advisory group’s biggest challenge likely lies ahead. When announcing the initiative, OpenAI casually dropped the news that its new, unreleased model—the same underlying system credited with resolving the Navier-Stokes Millennium Prize problem—has now resolved more than 100 long-standing open problems spanning most areas of mathematics. The company has previously signaled that it is also closing in on another Millennium Prize problem. On its official website, the advisory group noted that dealing with this immense logjam is its immediate priority, stating that it faces the very specific challenge of advising OpenAI on how to coordinate the release of a large number of significant mathematical results reported to have been produced by the internal model.
In part, this new effort at engaging mathematicians represents a tacit acknowledgment of the monumental public relations disaster generated by OpenAI’s previous announcements. During those earlier disclosures, the company found itself fiercely fending off accusations of scooping academic researchers, failing to credit the foundational human work its models built upon, and even alienating those who utilized its systems. These controversies led many mathematicians to wonder whether the tech company truly understood—or even respected—the rigorous discipline and tight-knit community it had barreled into.
Reflecting on the events of the past few months and the extensive commentary written about what mathematicians want, Kevin Buzzard, a mathematics professor at Imperial College London, expressed skepticism over why OpenAI felt it needed such an elite body simply to understand the community’s basic desires. It wasn’t proofs of hard theorems that people wanted, he argued, but rather a better understanding of the subject itself. It’s not entirely clear to me that you need a committee of brilliant people to hammer that point home, he said.
Leave a Reply