Mathematicians spend significant time manually verifying complex proofs and calculations. A tool could automatically generate human-readable verification certificates from AI outputs.
Create a collaborative interface where researchers can refine AI-generated proofs, with version control and peer review features.
Universities and research institutions would pay for time-saving tools that improve proof reliability.
Start with a simple web app that formats OpenAi's existing certificates clearly.
The biggest risk is AI still making subtle errors in complex proofs.