The AI takeover of mathematics has begun

Recent breakthroughs by artificial intelligence models, such as OpenAI's Astra and Anthropic's Claude, have successfully resolved several long-standing mathematical problems and conjectures that had previously stumped human researchers for decades.

While these achievements showcase the rapid advancement of generative AI in highly complex and abstract domains, they have also sparked intense debate within the mathematical community. Academics are grappling with a mixture of excitement over accelerated scientific discovery and deep anxiety regarding the future of their discipline, particularly concerning how these systems will alter the training and career prospects of future mathematicians.

Beyond career anxieties, the transition has highlighted ethical concerns regarding credit and commercialization. Prominent researchers have criticized AI companies like OpenAI for initially downplaying the essential human scholarship that laid the groundwork for their machines' proofs. This friction led to the drafting of the Leiden Declaration, which warns against corporate hype and urges stakeholders to protect human mathematical roles. While some believe AI will ultimately democratize research and free academics to focus on higher-level concepts, others worry the field will become lopsided, with AI churning out solutions without generating the new conceptual directions that human-led exploration naturally produces.