Technology news and discussion on TalkFuse.
Why I won’t leave mathematics, despite AI solving centuries-old problems | Ken Ono
Ken Ono says he nearly left mathematics after AI systems began tackling major problems, including a reported proof of the Riemann hypothesis. He argues that AI can expand what mathematicians can do, but people remain essential for judging whether a formal proof captures the intended mathematics and for understanding its meaning.
Ono now works at Axiom Math, where AI helped create a Lean formalisation of the classification of finite simple groups, producing 43.1 million lines of code. The project also uncovered a problem in a published argument and provided an alternative; the Lean checking system verified the formal proof. Ono recognises students’ concerns about credit, transparency and careers, and shares warnings from mathematicians about rushed announcements and loss of understanding.
In short
- Ono stayed in mathematics and now works on AI-assisted proofs.
- A formalisation of finite simple groups generated 43.1 million lines of Lean code.
- He says human judgement and clear attribution remain vital.
Steelman: both sides, in good faith
The strongest fair case each way — we don't pick a winner.
The case for
AI can take on laborious formalisation and checking, helping mathematicians test difficult arguments, find gaps and explore problems at a scale that would otherwise be impractical. If people remain responsible for interpreting results and teaching their meaning, these tools could extend mathematical work while leaving room for human judgement and creativity.
The case against
Mathematics depends on more than a proof passing a formal checker: researchers also need to understand why an argument works and what it means. If AI-generated results are announced faster than people can scrutinise them, the field could lose transparency, weaken learning and make credit and career prospects harder to judge.
Explained simply
Mathematics studies patterns and structures through logical arguments called proofs. Some proofs are so long or complex that checking each step is difficult; Lean is software that can verify whether a proof follows specified rules.
Ken Ono is a mathematician who works at Axiom Math, a company developing AI tools for mathematics. The classification of finite simple groups is a major result describing the basic building blocks from which certain groups can be assembled.
AI-assisted proofs raise questions about how mathematicians should check and explain results, and how credit should be shared. The tools may help researchers tackle work that would be hard to manage by hand, while people still need to assess what a proof establishes and what it means.