Science
AI Is Helping Mathematicians Solve Problems No Human Could Check Alone
What if a maths proof had a million steps — how would anyone know it was right? Terry Tao, the world's most celebrated living mathematician, is now championing AI tools that can check every single step automatically.
hundreds of pagesLength of some modern maths proofs that AI tools can now help verify
The facts
- 1Terry Tao, a Fields Medal winner often called the greatest living mathematician, has publicly backed using AI proof-checkers to help solve extremely complex problems in mathematics.
- 2Automated proof-checkers work by breaking a big maths problem into thousands of tiny logical steps, verifying each one independently, then reassembling them — like checking every brick in a wall before trusting the whole building.
- 3Some modern proofs are so long they run to hundreds of pages, making human error nearly impossible to avoid; AI tools can scan these with a consistency no single person can match.
- 4Tao and collaborators have begun using a system called Lean, a formal verification tool, to confirm that key parts of difficult proofs are logically airtight — a process once considered too slow for real research.
- 5Critics note that AI proof-checkers currently require mathematicians to rewrite proofs in a special computer language first, which takes significant extra time, so the tools help most on big collaborative projects, not quick everyday calculations.
Why it matters
Mathematics is the foundation of everything from coding to engineering to medical research. If AI can reliably verify complex proofs, it could unlock problems that have stumped humanity for decades — but it also raises questions about whether mathematicians of the future will need different skills than those who came before.
Sources
- Quanta Magazine
- Fields Medal Committee (International Mathematical Union)
