Mathematicians Warn Over AI Attribution Disputes as OpenAI’s Altman Rules Out 2026 IPO
aiinsider ·
Twenty-five Fields Medal-winning mathematicians signed an open letter cautioning that AI laboratories are endangering mathematical research by rushing to solve notable problems without correct attribution or verification. The letter came after NYU professor Tristan Buckmaster accused OpenAI of pressuring him not to credit an Anthropic-affiliated collaborator on a major proof, and after speculation that OpenAI may have drawn on data from his Codex sessions for a competing proof. OpenAI then withdrew its sponsorship of a Caltech math event amid objections from researchers there. The signatories said AI-generated solutions are frequently announced too quickly, lacking adequate writeups and citations to prior work, which raises attribution and plagiarism concerns and could lead mathematicians to favor secrecy over open collaboration. This letter followed a comparable June effort called the Leiden Declaration. In a separate development, OpenAI CEO Sam Altman told Fortune’s Alyson Shontell that the company will not seek an IPO in 2026, even though it had filed confidentially for one. Altman cited current safety issues, including the fallout from a recent security incident involving Hugging Face, as reasons the timing was poor, and said OpenAI would only move forward when it judged the business and wider societal context ready.
aiinsider