OpenAI published more than 350 mathematical discoveries generated by artificial intelligence, prompting top researchers to react and warning of a transformation for the discipline. Alex Kontorovich, the mathematics department chair at Rutgers University, stated that if a human had produced one of the showcased proofs, it would undoubtedly earn a Fields Medal. The release includes tens of thousands of pages of results addressing open questions across algebra, number theory, theoretical computer science, mathematical logic, and topology.
Oxford Mathematician Calls AI Mathematical Output Impressive
Martin Bridson, an Oxford University mathematician and president of the Clay Mathematics Institute—the organization that established the Millennium Prize Problems—described the OpenAI publication as impressive. Bridson noted that until recently, imagining such extensive movement across the frontiers of mathematics in a single day was impossible. The Clay Mathematics Institute president added that the solutions announcement will prove exciting for many, terrifying for others, and devastating for some. Bridson cautioned that researchers must embrace machine capabilities without becoming subordinates who merely follow their lead.
Quasi-Riemann Hypothesis Proof Generates Shock and Debate
None of the proofs completely resolve the five remaining Millennium Prize Problems, but the findings include indirect advances connected to all of them. The most significant result motivating Kontorovich’s reaction is a proof of the quasi-Riemann hypothesis, a conjecture directly linked to the Riemann hypothesis that could serve as a preliminary step toward a full resolution. Ken Ono, a University of Virginia professor and founder of the AI mathematics firm Axiom Math, stated that many mathematicians experienced a physical sensation akin to being punched in the stomach upon reviewing the findings. Ono added that this initial shock is quickly yielding to enthusiasm and determination as researchers recognize the scale of the new mathematical territory.
Decades-Old Conjectures and Rokhlin Problem Findings
Bryna Kra, a mathematician at Northwestern University, focused on problem number 145 on the list, which addresses Rokhlin’s problem regarding whether the mixing property implies higher-order mixing—a question originally posed in the 1940s. Kra called the results impressive while emphasizing the need for careful human review of the tens of thousands of generated pages. Meanwhile, French mathematician Jean-Pierre Serre, an emeritus professor at the Collège de France and recipient of both the Fields Medal and the Abel Prize, found particular interest in result number 46, which confirms a conjecture he formulated nearly seventy years ago. Serre noted that he utilizes AI for tasks like finding references and investigating intuitions, but expressed concern that heavy reliance on automated discovery could diminish the deep pleasure of learning and discovery, particularly for younger mathematicians.
Related reading
