OpenAI's announcement in early September that it had solved a long-standing mathematical problem sent shockwaves through the global mathematics community, prompting immediate debate about the future role of human researchers. The claim — that a difficulty which had resisted solution for almost a century had been resolved by an AI system — left many mathematicians reeling and searching for clarity about the provenance and verification of the result.
Reaction among academics was swift and mixed. Some praised the technical achievement while others voiced deep unease, questioning how the discipline should respond when machine systems can tackle problems that have defined generations of human work. A number of mathematicians characterised the company’s public presentation of the result as operating in an "immature playground" of boasting rather than in the cautious, peer-reviewed tradition that typically governs major advances in the field.
The unease culminated in an open letter signed by 25 recipients of the Fields medal, often referred to as the Nobel prize of mathematics, expressing fear of what they described as a "severe misalignment" between the priorities of AI companies and the needs of mathematical research. The letter highlights broader anxieties about oversight, standards of proof, and the social and professional consequences for researchers as AI systems become increasingly capable of resolving complex problems.
To explore how these developments might reshape the discipline, Guardian science editor Ian Sample spoke with Colva Roney-Dougal, a professor of pure mathematics at the University of St Andrews. Their discussion examined the ways in which advanced AI tools could both accelerate discovery and challenge existing norms around verification, explanation and the training of new mathematicians.
The conversation and the wider debate underline a pivotal moment for mathematics: the community must now consider how to integrate powerful computational tools while preserving the methodological rigour and communal trust that underpin the discipline. As researchers, institutions and companies grapple with these questions, the longer-term consequences for who does mathematics and how it is done remain unresolved and are the subject of intense ongoing discussion.