While the global community fears that AI will replace humans in various fields, the AMU Artificial Intelligence Centre has just proven that mathematicians can rest assured. D.Dr Bartosz Naskręcki, the only Polish member of an international team of 30 mathematicians, created a test that completely defeated the latest AI models. Dr Naskręcki worked with experts in number theory and algebraic geometry, including Prof. Ken Ono, one of the world's leading experts in number theory, and Prof. Ravi Vakil, a leading authority in algebraic geometry and currently also president of the American Mathematical Society.
AI Surrenders to Real Mathematics
The results are devastating for artificial intelligence. The world's best AI model, OpenAI's o4-mini, managed to solve only 6.3% of the most challenging mathematical problems. It is as if a student who knew only the multiplication table were trying to pass an exam in mathematical analysis. "This shows that AI is still like a very advanced calculator – it can perform complex calculations, but it does not understand deep mathematics," explains Dr Naskręcki. "True mathematical reasoning requires creativity, intuition and the ability to connect seemingly unrelated concepts – something that machines are still unable to do."
Mathematicians are Still Irreplaceable
The FrontierMath project, in which Dr Naskręcki participated, is the most advanced mathematical test for AI in history. It consists of 350 problems that have been specially designed to test the limits of artificial intelligence. The result? AI failed spectacularly.
Most of the latest models – including Google DeepMind, Anthropic and xAI systems – have not solved a single problem of the highest level of difficulty. It is as if the best chess players were asked to solve a problem that no human has ever solved.
Experts predict: AI will not threaten mathematicians for decades
Terence Tao, one of the world's most prominent mathematicians and a Fields Medal winner, is categorical: "These problems will resist AI systems for several years."
Timothy Gowers, also a Fields Medal winner, adds: "This is a completely different level of difficulty than anything AI has been able to solve so far."
Prof. Igor Pak of UCLA goes even further, predicting that some problems may remain unsolved by AI for the next 50 years: "This shows that mathematics is not just about calculations – it is the art of thinking" he emphasises.
Polish Success on the World Stage
Dr. Naskręcki, the sole representative from Poland and one of only five mathematicians from Europe, was part of an elite group tasked with creating a groundbreaking test. "This is proof that Polish mathematicians remain at a world-class level and that our skills cannot be replaced by machines," he proudly declared.
The AMU Artificial Intelligence Centre, where Dr Naskręcki is a member of the Quantum Informatics Team, is thus becoming one of the key research centres on the frontiers of AI, not to exceed them, but to discover where these boundaries lie.
The Future Belongs to Humans
While the media is full of stories about how AI will replace doctors, lawyers and programmers, mathematics remains a bastion of human intelligence. FrontierMath confirms that true scientific thinking, creativity and problem-solving skills are still the domain of humans.
"AI can be a tool, but it never replaces mathematics. It is like a hammer replacing an architect. It can help with construction, but it can not design a house," concludes Dr Bartosz Naskręcki.
photo: Piotr Jabłoński