OpenAI's 370 math breakthroughs spark verification crisis
OpenAI released 370 mathematical results including solving the Navier-Stokes equation, but experts say AI proofs can't be verified by humans.
OpenAI published over 370 new math results this week across algebra, computer science, and logic. Last month the company solved the Navier-Stokes equation, one of the world's hardest unsolved problems (worth US$1 million to whoever cracked it).
The Institute for Advanced Study in Princeton said it does not endorse the practice. The core issue: AI can write math arguments that no human can understand, check, or take responsibility for.
The broader math community cannot access these models. It cannot independently check the work or build on it.
OpenAI says it will work with the Institute going forward but has not committed to stopping. Experts worry the lab is not properly vetting results.
- 370
- Mathematical results published
- US$1 million
- Navier-Stokes reward
Why it mattersIf AI can solve hard math problems no human can verify, who gets credit and how can the field trust the answers?
AustraliaNo direct Australian effect yet, but if proprietary AI labs monopolise mathematical discovery, Australian researchers could lose access to tools and insights the broader field depends on.
✓ Claims checked against the source and corrected before publish. checked 2 h ago
Open this story in InSnip →
