Mathematicians want proof OpenAI didn’t use their work

A dispute is currently unfolding involving mathematicians and the artificial intelligence companies OpenAI and Anthropic. The situation describes a mathematician who feels caught or crushed between the two competing AI entities as they navigate the complexities of mathematical proof and training data. Coverage from The New York Times emphasizes the tension between the individual contributors of mathematical theory and the corporate scale of the AI industry. The report highlights a specific instance where a math problem serves as the catalyst for this broader disagreement.

By naming both OpenAI and Anthropic, the coverage suggests that this is not an isolated incident with a single company, but rather a systemic issue affecting multiple leading developers of large language models who are increasingly integrating advanced mathematical reasoning into their systems. The context of this trend involves the intersection of academic mathematical research and the commercial development of AI. As AI companies strive to improve the logical and mathematical capabilities of their models, they often rely on high-level proofs and academic papers. The matter is significant now because it raises fundamental questions about ownership and the boundaries of fair use when proprietary or academic mathematical insights are used to train commercial software that may eventually automate the very tasks the mathematicians performed.

Moving forward, the situation will likely center on whether the mathematicians can obtain definitive proof regarding the training sets used by OpenAI and Anthropic. The New York Times report indicates that the mathematician in question is currently positioned between these two entities, suggesting that further revelations regarding the specific math problem and the internal data practices of these AI companies may emerge. Observers will be watching to see if this leads to formal demands for transparency or legal challenges regarding the use of mathematical work.