Mathematicians are demanding that OpenAI provide proof it did not use their copyrighted work to train its AI models. The controversy follows recent revelations that OpenAI's models can reproduce mathematical proofs and theorems with high fidelity, suggesting possible memorization of training data. The mathematicians argue that without transparency, it is impossible to verify whether their intellectual property was used without permission. OpenAI has not publicly responded to the specific demands, but the company has previously stated that its models are trained on publicly available data and that it respects copyright. The dispute highlights growing tensions between AI developers and content creators over the use of copyrighted material in AI training.
This is not just another copyright spat. Mathematicians are the canaries in the coal mine. Their work is precise, structured, and often unique. If OpenAI's models can spit out their theorems verbatim, that's not inspiration. That's copying. And mathematicians know it.
I believe in AI's potential to augment human knowledge, not replace it. But that future requires trust. And trust requires transparency. OpenAI should open its books. Not just for mathematicians, but for all of us. Because if we can't verify how models are trained, we can't verify what they'll do next. This is a chance for OpenAI to lead. To show that AI can evolve with ethics, not despite them. Let's hope they take it.