OpenAI proof dispute raises concerns about trust in AI research
A public dispute has emerged over an AI-generated result concerning the Navier–Stokes equations, one of the Clay Millennium Problems carrying a $1 million prize. Mathematician Tristan Buckmaster says OpenAI used material from drafts he prepared with Levent Alpöge, pressured him, and sought to exclude Alpöge from authorship because he works for Anthropic. Buckmaster described the conduct as “absolute academic malpractice.”
OpenAI CEO Sam Altman and researcher Sébastien Bubeck rejected plagiarism allegations. Altman said the effort began after online rumors that Anthropic models had solved a Millennium Problem. OpenAI employees argued that training on the researchers’ submissions was unlikely, while the company’s own statement said it could not rule out that de-identified usage data had helped improve its models. Alpöge disputed Altman’s account of their communication and said he would have welcomed cooperation.
The episode has prompted broader concerns about open science. Terence Tao warned that even rumors could lead technology companies to deploy substantial resources and overtake original research projects. He said this could discourage researchers from sharing promising ideas and damage long-standing traditions of scientific openness.
Written by Sakhanda Wire AI from reporting by The Decoder. Facts, quotes and full context are in the original article below.