344
Productivity & Workflow355
Automation & Workflow225
Software Development251
Marketing & Growth193
AI Infrastructure & MLOps175
Writing & Content Creation204
Data & Analytics142
Photography & Imaging156
Design & Creative170
Customer Support133
Sales & Outreach125
Voice & Speech135
Education & Learning131
Operations & Admin87
Mathematicians say AI-made proofs can be hard to verify and understand. OpenAI is creating a group of experts to advise on its math work.
In short: Mathematicians are raising concerns that AI can produce math proofs faster than people can understand or check them, and OpenAI has responded by forming an advisory group.
OpenAI recently said it used “thousands of agents” to tackle the Navier-Stokes existence and smoothness problem, a famous math puzzle about equations that describe how fluids flow. The company published a 166-page proof, but it is still under peer review, meaning other experts have not yet confirmed it is correct.
Several mathematicians say the bigger issue is not just whether the answer is right, but whether people can follow how the proof was made. In normal math research, a proof is like a carefully written recipe. It shows each step and points to earlier work so others can repeat and trust it.
Researchers quoted by Wired say AI systems often do not clearly explain their steps or reliably cite sources. Some also raised concerns about transparency and attribution, including suggestions that earlier human work may have been used without proper credit.
A group of mathematicians signed an online statement called “A Severe Misalignment of AI in Mathematics.” It warns that mass-producing proofs could harm the parts of math that depend on slow learning and deep understanding, especially for students and early-career researchers.
OpenAI says it has formed an advisory group of mathematicians to guide how it uses AI in mathematics. The next questions are whether the Navier-Stokes proof holds up under review and whether AI companies will share enough detail for the wider community to verify results and learn from them.
Source: Wired