OpenAI forms math advisory group as its AI resolves more than 100 open problems

Summary

Following claims that its AI solved the Navier-Stokes Millennium Prize problem and over 100 open problems, OpenAI established an advisory group at the Institute for Advanced Study (IAS) in Princeton. However, this group was explicitly prohibited from influencing the speed or direction of the company's mathematical research.

Why it matters

This move comes amid a public reckoning. Twenty-five Fields Medalists signed a letter accusing AI labs of harming mathematics through a competitive arms race. By establishing a group with no decision-making authority or say in research speed, OpenAI is managing reputational risk while retaining full operational control over its mathematical workflows.

Highlights

• The group has zero decision-making power and will not advise on the pace of OpenAI's internal mathematical research.
• Only one of the first nine members (Camillo De Lellis) also signed the critical open letter by the 25 Fields Medalists.
• OpenAI claims that a single internal model solved over 100 open problems in 'most areas of mathematics'.

Critical take

This structure is telling: members are voluntary, have no veto power, no authority to set the pace, and the host, IAS, explicitly waives any influence over the company. This reads less like genuine community governance and more like a PR shield designed to absorb backlash while research continues unabated.