AI-generated mathematical results are arriving quickly enough to challenge a basic assumption about science: that new discoveries come only from human minds. On Tuesday, Hungary’s Minister for Science and Technology said the country should help shape this shift rather than simply follow it.
In a post on his Facebook page, Tanács Zoltán cited an OpenAI publication dated October 6. It contained more than 700 mathematical research manuscripts, organized into 372 result clusters. The post said one of the world’s leading AI companies had used a new, unpublished AI model to produce new mathematical results in hundreds of cases.
Each result required an average of just three hours of computing capacity equivalent to ChatGPT Pro, according to the post. The phenomenon has been dubbed “Mathocalypse” (“Mathematical apocalypse”), a name that reflects the possibility that the era of treating human thought as the sole source of new scientific discoveries is ending.
The cited results span four areas: questions related to understanding the distribution of prime numbers; progress on long-studied problems involving the volumes of high-dimensional bodies; counterexamples to decades-old mathematical conjectures in algebra; and new findings on how accurately pi can be approximated with ordinary fractions.
A substantial portion of the proofs has also been formalized in Lean, a mathematical system that lets machines check their logical correctness. But the published results are still undergoing independent scientific review, the post cautioned, so not every problem can yet be considered definitively solved.
For centuries, leading scientists could spend years, sometimes decades, working on a single problem. Now, a technology can produce new scientific results in as little as a few hours. Takács Zoltán argued that this won’t make people redundant: “I am convinced that humans will not become obsolete. But our role will fundamentally change.”
In his view, being the fastest to find a solution will matter less. What will matter more is asking the right questions, understanding and checking results, and using them responsibly. Takács said keeping human oversight is essential.
That shift calls for more than making existing education, research, and innovation systems slightly more efficient. Takács said the way scientific research and innovation work needs to be rethought. For Hungary, he described AI as both a historic opportunity and a major responsibility: it could give a smaller country access to scientific and innovation capabilities once available only to the world’s wealthiest institutions.
