AI integration in mathematics sparks academic integrity concerns

The mathematics community is facing growing tension over the role of artificial intelligence in solving complex, open research problems. At the center of the debate is OpenAI, which recently announced that an unreleased internal model produced a possible solution to one of the seven Millennium Prize Problems in just four days.

This development has raised concerns regarding intellectual property and academic integrity. Mathematician Tristan Buckmaster expressed concern that OpenAI’s model may have utilized his unpublished research on the Navier–Stokes equations after he input drafts into the company’s Codex tool. While OpenAI denied the work influenced the system, the company noted that user interaction data may be used to improve its models.

Simultaneously, a planned “Mathathon” at the California Institute of Technology (Caltech) faced significant backlash from faculty and researchers. Critics argued the event, originally sponsored by Anthropic and OpenAI, would promote “slop mathematics” and serve as corporate public relations. In response, student organizers redesigned the competition to focus on developing alternative proofs for existing solutions rather than tackling open problems, and removed the AI companies as sponsors.