Mathematicians divided over AI’s research upheaval

OpenAI’s solution to the Navier-Stokes equations has intensified debate over AI’s role in mathematics. The company subsequently released solutions to an additional 700-plus math problems on GitHub. NYU professor Tristan Buckmaster said the release wiped out entire research programs. Sam Altman has suggested mathematicians are moving from anger toward acceptance, but prominent researchers remain cautious. UCLA professor Terence Tao said exciting possibilities are emerging while mathematics is losing elements of its “romantic era.”

The tensions have spilled into a student-run Caltech event allowing participants to use AI. Organizer Caiman Moreno-Earle described the rapid breakthroughs as demoralizing and framed the “mathathon” as an effort to preserve a place for humans in mathematical research. Anthropic and OpenAI each initially committed $1 million worth of computing power. A critical letter attracted more than 700 signatories and warned of destructive consequences for the mathematical community. OpenAI subsequently withdrew, although the event remains scheduled.

Beyond concerns about disrupted careers and research priorities, mathematicians face practical questions about reliability and ownership. AI-generated solutions can lack clarity, leaving researchers to verify the work. Some mathematicians also speculate that ideas shared while brainstorming with models could be used to reach solutions ahead of them, adding uncertainty about how these systems produce their results.