OpenAI Forms Elite Mathematics Advisory Panel After Months of Reputational Friction

After turning a string of spectacular mathematical results into a severe reputational crisis, OpenAI is attempting to chart a less disastrous path forward by consulting human mathematicians. The artificial intelligence titan announced the creation of an independent panel of mathematicians tasked with advising both OpenAI and other AI companies on their interactions with mathematical research and the wider academic community. The group’s abrupt arrival caught many researchers by surprise, sparking widespread discussion about its mandate, its level of influence, and whether such a small and elite group can truly represent the collective interests of the global mathematical community.

The newly formed advisory body comprises nine prominent researchers drawn from prestigious institutions such as Stanford, Harvard, Oxford, and Cambridge. Among them are multiple Fields Medalists and recipients of MacArthur “genius grants.” The group will be hosted by the Institute for Advanced Study (IAS) in Princeton, New Jersey, one of the world’s most revered centers of mathematical research and the historical intellectual home of legendary figures including Albert Einstein, John von Neumann, and J. Robert Oppenheimer.

OpenAI stated that the organization, officially named the Advisory Group on Mathematics and Artificial Intelligence (AGMAI), will operate independently. According to the company, the panel will have the freedom to offer unsolicited advice, publicly comment on the impact of AI on mathematics, and make its guidance public, though specific operational details remain scarce. It appears the group will be granted early access to OpenAI’s research to fulfill its responsibilities. Members will not receive financial compensation from OpenAI, and the group retains the authority to modify its membership as it sees fit. OpenAI emphasized that the value of the panel relies on its members exercising independent judgment and challenging company practices. The group is expected to advise on the review and communication of emerging results, help assess their significance, coordinate dissemination, and uphold academic and professional standards. Additionally, it will weigh in on how AI tools can support mathematical research and learning.

Despite OpenAI’s emphasis on institutional independence, many researchers remain skeptical about the group’s actual influence and the obscure circumstances surrounding its formation. Simon Machado, a researcher at ETH Zurich in Switzerland, remarked that the initiative appeared shrouded in mystery and questioned the selection process for the nine members. The group’s bare-bones website indicates that it came together after OpenAI approached certain individuals about establishing an external advisory board. Following discussions with the company, those mathematicians decided instead to launch an independent group and invite others to join. It remains unclear which members were initially approached by OpenAI or how the remaining participants were chosen.

Addressing what he termed a torrent of misinformation surrounding the panel’s inception, advisory group member Martin Hairer published a blog post elaborating on its origins and underlying philosophy. Hairer noted that AI has already exerted a profound impact on mathematical research, raising complex questions regarding the correct attribution of ideas, priority, and human understanding. He argued that the mathematical community cannot afford to adopt an ostrich approach by simply ignoring AI laboratories and refusing to dialogue with them on principle.

While acknowledging that the group was formed in response to an initial outreach from OpenAI, Hairer stressed that not all members were originally contacted by the company and that the resulting organization operates independently of any frontier AI lab. Beyond standard confidentiality requirements, members have signed no restrictive agreements governing their public statements. They receive no financial compensation and rely solely on the IAS for logistical and technical support. Hairer emphasized that the group’s overarching aim is to represent the interests of the mathematical community as effectively as possible. While conceding that a nine-member panel is hardly a representative statistical sample of the entire field, he expressed absolute certainty that all participants are acting in good faith.

Nevertheless, the elite academic composition of the panel has prompted questions about whether it possesses the broad range of expertise necessary to evaluate everything crossing its path, given the highly specialized skills required at the frontiers of modern mathematics. Francesco Fournier-Facio, an incoming mathematics professor at Heriot-Watt University in Scotland, expressed ambivalence about the panel. He suggested that greater engagement with mathematicians is inherently positive, but questioned whether the group reflects the broader community whose work and livelihoods are being disrupted by OpenAI’s quest for high-profile mathematical achievements. Fournier-Facio described the initiative as having an ivory tower dynamic.

Machado echoed these sentiments, questioning whether the panel’s members fully understand or can speak to the concerns of everyday mathematicians regarding artificial intelligence. He noted that while the members are extraordinary mathematicians, their elite reality differs drastically from the day-to-day circumstances most working mathematicians face.

Earning the trust of the wider mathematical community will ultimately determine whether the advisory group can successfully mitigate the fallout from OpenAI’s recent missteps in announcing results and cushion the impact of future announcements. In his blog post, Hairer acknowledged the risk that AI labs might attempt to spin the group’s statements to suit corporate public relations machines that dwarf academic resources. Nevertheless, he argued that refusing to engage now would be hypocritical, particularly after widespread criticism of OpenAI’s previous conduct and communication failures. While the group has not yet finalized its specific recommendations, Hairer pointed out a clear and acknowledged problem: AI companies have recently produced high-profile mathematical results, but their publication and dissemination have fallen far short of acceptable academic standards.

For the immediate future, the advisory group is actively gathering input through colleague discussions, a public feedback form, and engagement on online forums like Proofs and Prompts, which has become a vital meeting ground for mathematicians grappling with technological disruption.

The panel’s most formidable challenge, however, lies immediately ahead. When announcing the advisory group, OpenAI casually revealed that a new, unreleased model—the same system credited with resolving the Navier-Stokes Millennium Prize problem—has now resolved more than 100 long-standing open problems across nearly all areas of mathematics. The company has also previously indicated that it is closing in on another Millennium Prize problem. The advisory group noted on its website that addressing this monumental logjam is its absolute priority, describing its current challenge as advising OpenAI on how to coordinate the release of a massive volume of significant mathematical results generated by its internal systems.

This outreach effort largely acknowledges the public relations disaster that unfolded when OpenAI was accused of scooping researchers, failing to properly credit human foundational work, and even utilizing the intellectual property of those interacting with its models. These controversies led many mathematicians to question whether the company respected the academic discipline it had entered. Kevin Buzzard, a mathematics professor at Imperial College London, observed that after months of extensive public discussion regarding community desires, it remains questionable whether an elite committee is truly necessary to understand what mathematicians want. Buzzard argued that the fundamental issue was never a lack of proof for hard theorems, but rather a deeper understanding and respect for the subject matter, suggesting that a brilliant committee may not be required to state the obvious.