The speed at which artificial intelligence is gaining mathematical capabilities has taken many by surprise. As New Scientist noted in an article, the “new” world is rewriting what it means to be a mathematician.
In March 2025, the mathematician Daniel Litt made a bet. Despite the march of progress in artificial intelligence in many fields, he believed his discipline was secure, betting with a colleague that there was only a 25% chance that an AI could write a mathematical paper at the level of the best human mathematicians by 2030.
Just a year later, he thinks he was wrong. “Now I hope to lose this bet”, he declared, quoted by .
Mathematicians are increasingly surprised by the speed of improvements in AI’s ability to solve problems and produce demonstrations.
“A couple of years ago, they were basically useless even for solving high school math problems, and now they can sometimes solve problems that actually appear in a mathematician’s research life,” says Litt of the University of Toronto.
The game has changed and This progress is faster than many had predictedwith mathematicians warning that their profession is undergoing one of the fastest evolutions the field has ever seen.
“We’re running out of places to hide. We have to face the fact that the AI will soon be able to prove theorems better than us“, he wrote Jeremy Avigadfrom Carnegie Mellon University, in Pennsylvania, in a recent essay, which the same magazine cites.
Companies like OpenAI and Google DeepMind have already achieved gold medal performances in the International Mathematics Olympiad, an elite competition for high school students that many experts had previously considered to be beyond the reach of AI tools.
AI has begun to tackle more complex mathematics, solving real research problems and helping to automatically verify cutting-edge demonstrations, which traditionally could require a huge amount of work on the part of teams of mathematicians.
“Things changed very quickly,” he says. Thang Luongda Google DeepMind.
AI company Math, Inc. recently surprised mathematicians by announcing that its AI tool, called Gausshad formalized an award-winning demonstration and verified that it was correct. The demonstration concerned how many spheres can be packed into a space and was the subject of Maryna Viazovska’s 2022 Fields Medal, often called the Nobel Prize of mathematics.
As New Scientist details, the effort to formalize Viazovska’s work began with a small group of mathematicians in late 2024, working separately from Math, Inc., led by Sidharth Hariharanfrom Carnegie Mellon University, who hoped to manually translate the problem into computer code. First they analyzed Viazovska’s sphere packing solution in eight dimensions. While making steady progress, Math, Inc., which later assisted the researchers, announced that it now had a complete proof and then a more general version of a result for 24 dimensions.
A new era of mathematicians
“The future we are all thinking about is that we will have tools that will automatically formalize new investigations and mathematical papers, and point out whether there are errors or not. This will have huge implications for, for example, peer review and assessment work,” he said Johan Commelinfrom the University of Utrecht, in the Netherlands.
Facing a future in which an increasing share of mathematics is done by AI, some mathematicians, like Avigad, are sounding the alarm about the harmful effects this could have on our ability to practice and create new mathematics.
Mathematicians nevertheless remain hopeful that there will be a place for them in an increasingly machine-led future.
Looking back at history, Commelin says that manual calculations were once a big part of being a mathematician, but are now done automatically: “Similar things will happen here, where we will radically change what we do, but in 10 or 20 years we will continue to recognize what we do as mathematics, in a new style.”
