OpenAI announced on Monday a new independent Advisory Group on Mathematics and Artificial Intelligence, hosted at the Institute for Advanced Study (IAS) in Princeton, New Jersey — and revealed that the same internal model behind its contested Navier-Stokes result has, by its own count, resolved more than 100 additional open problems across most areas of mathematics.
The group is meant to give mathematicians a formal channel into the company's math-oriented research at a moment when the research community is openly frustrated with how those results are being released. "This group will serve as a bridge to the mathematical community and broader public, giving mathematicians a voice in how we move forward," OpenAI said in its announcement, as quoted by TechCrunch. For more context on this story, see our ongoing latest AI developments.
A Response to the Navier-Stokes Backlash
The advisory group's creation follows the abrupt publication of a claimed solution to the Navier-Stokes Millennium Prize problem, a result that stunned mathematicians both for its ambition and for the pace of its release. In practice, OpenAI's new group will mostly serve an advisory role: assessing the significance of new results and coordinating their release, so that purported breakthroughs arrive with context and expert scrutiny rather than as surprise announcements.
The scrutiny has not let up. Scientific American published a piece on Monday asking "Did OpenAI solve the wrong Navier-Stokes problem?" — a reflection of how much uncertainty still surrounds the claim even among science journalists tracking it closely.
Many prominent mathematicians have taken issue with the frenzied pace of recent claims more broadly. Earlier this month, 25 Fields Medal-winning mathematicians signed an open letter arguing that AI labs are threatening mathematicians' intellectual work as they compete to one-up each other with solutions to famous problems. The concern is not that the problems get solved — it is that a centuries-old culture of careful, verified publication is being steamrolled by labs racing for headlines.
Advisory Power — With Hard Limits
The group's structure attempts to balance independence with OpenAI's clear intention to keep control of its research agenda. Members will not be paid. They can offer unsolicited advice, go public with their views and control their own membership — provisions that give the body a measure of independence from the company that created it.
But the group has no power to slow anything down. As OpenAI's post puts it, "the group will not be responsible for advising us on how to pace our internal progress on mathematics."
The Institute for Advanced Study was emphatic on the same point in its own announcement. "Although we will give advice, we do not have decision making power at any AI company, and the responsibility for the decisions made by any company will rest with that company," the institute said in a statement.
That division of labor — advice without authority — is the compromise that made the partnership possible. IAS gets a seat at the table for mathematics; OpenAI keeps the steering wheel.
Nine Members, Little Overlap With the Critics
Nine prominent mathematicians have been named as initial members of the group. Notably, only one of them — IAS's Camillo De Lellis — also signed the Fields Medalists' letter earlier this month, a sign that OpenAI has not simply recruited its loudest critics into the room.
Whether that matters depends on what the group does with its platform. Its members can publicly disagree with the company, and the design invites them to. An advisory body that never voices dissent would settle the question of its value quickly.
Why It Matters
The announcement is the latest flashpoint in a debate about how AI laboratories should handle machine-generated scientific claims. A model claiming to settle a Millennium Prize problem is an extraordinary assertion, and the mathematical community's core complaint is procedural: results of this magnitude deserve verification and measured release, not press-cycle announcements.
Verification is the real bottleneck. A proof that takes a model days to produce can take human experts months to check, and mathematics has no peer-review infrastructure designed for that asymmetry. That is precisely the problem the group's mandate — assessing significance and coordinating release — is built to address, at least in principle.
OpenAI's move acknowledges that criticism while stopping well short of conceding control. The group can lend legitimacy, coordinate how results are presented and give outsiders a public platform — but the company retains full authority over what its models pursue and how fast. The claim of 100-plus solved problems now sits with a body that can shape how it is announced, not whether it is made.
Whether mathematicians treat the advisory group as a meaningful check or a fig leaf will depend on what happens next: whether the next major claim arrives with the group's imprimatur, and — more importantly — with proofs the community can actually verify.
---
Stay Ahead of AIGet the latest AI news, analysis, and breakthroughs — all in one place.
Read more AI news →
