The AI takeover of mathematics has begun

Key Points:

• OpenAI’s advanced AI model Astra recently solved 10 long-standing mathematical problems, demonstrating AI’s growing capability to accelerate mathematical discovery by combining existing knowledge in novel ways.
• The announcement sparked mixed reactions in the mathematics community, with excitement about AI’s potential tempered by concerns over credit attribution, the role of human mathematicians, and the future of the discipline.
• Some mathematicians criticized OpenAI for downplaying prior human contributions, particularly in the case of non-sofic groups, highlighting tensions around recognition and the narrative promoted by AI companies.
• There are worries about AI’s impact on academic funding, access disparities, and the development of future mathematicians, as AI may solve problems typically tackled by graduate students, potentially undermining traditional training and career prospects.
• While some researchers remain optimistic about AI broadening participation and freeing humans for more creative work, many express uncertainty and fear over the rapid changes and the long-term implications for the field’s structure and values.