From Problem-Solver to Discoverer
Artificial intelligence has long been an indispensable tool in science, prized for its ability to analyse massive datasets faster than any human could. But a significant shift is underway. AI is evolving from a high-speed data cruncher into a creative
partner capable of making novel scientific discoveries. This evolution reached a stunning new peak with recent breakthroughs in mathematics, a domain long considered the pinnacle of human abstract reasoning. Instead of just solving problems we set for it, AI is beginning to tackle longstanding, unsolved challenges that have stumped mathematicians for decades. In some cases, AI models are generating proofs and solutions that are not only correct but use strategies that humans had never considered, marking a new chapter in the relationship between technology and science.
The Maths Problems AI Is Cracking
The breakthroughs are not just theoretical. In one recent example, an AI model from OpenAI disproved a long-held conjecture related to the 'unit distance problem,' a famous question first posed by mathematician Paul Erdős in 1946. For nearly 80 years, the prevailing belief about the problem was unchallenged, but the AI autonomously found a new construction that provided a better solution. This wasn't a one-off event. Google DeepMind's FunSearch system has also made new discoveries in complex areas like the 'cap set problem'. FunSearch works by pairing a creative large language model (LLM) that generates solutions in the form of computer code with a rigorous evaluator that weeds out incorrect ideas, effectively mimicking an evolutionary process of discovery. This shows AI is not just finding answers, but generating the very process of how to find them.
Why This Is a Milestone
The significance of these achievements lies less in the specific problems solved and more in what they represent for the future of science. For the first time, AI models are autonomously resolving prominent open problems at the heart of entire mathematical subfields. Unlike previous computer-assisted proofs where humans guided the process, these AI systems are demonstrating a form of creative reasoning. They are uncovering unexpected connections between different fields of maths and pushing existing techniques to their limits. Experts describe this as a fundamental shift, transforming AI from a tool into a collaborator. This new partnership has the potential to dramatically accelerate the pace of discovery, not just in mathematics but in fields like drug discovery, materials science, and climate modelling that rely on complex mathematical foundations.
A New Kind of Scientific Partner
The rise of the AI mathematician doesn't mean human scientists are becoming obsolete. Instead, it signals the beginning of a new era of human-AI collaboration. Mathematicians who have verified AI-generated proofs note that while the core solution may come from the model, human insight is still crucial for discussing, refining, and understanding the deeper implications of the discovery. The AI can be seen as a tireless research assistant or a 'colleague' with an alien-like intelligence, capable of spotting patterns that human minds might miss. This allows scientists to focus on the bigger picture: asking the right questions, interpreting the results, and determining which problems are most important to solve. The goal is a symbiotic relationship where AI handles the immense computational and pattern-finding load, freeing up human researchers to do what they do best: exercise intuition, abstraction, and genuine understanding.
