Mathematicians want proof OpenAI didn’t use their work
The Verge Robert Hart ● Covered by 21 sources
A second mathematician is accusing OpenAI of using his work without being clear about it. The fight is now about where the company’s math breakthroughs really came from.
Based on reporting by The Verge, Robert Hart — read the original for the full story.
Summary, retelling and take written by AI under human oversight; images are AI-generated illustrations. How we work · Report an error
OpenAI’s math gains are under more scrutiny, and the questions are getting sharper. Just days after one dispute over whether the company’s models benefited from unpublished work, another mathematician has stepped in with a fresh complaint about how OpenAI handled its training data.
This time the critic is Andreas Thom. In posts on Mastodon, he accused the company of unethical and “dishonest” behavior and said it had not been transparent about where its training data came from.
Thom’s concern is narrower than a general gripe about AI scraping the internet. He says the interactions he and his colleagues had with ChatGPT before OpenAI’s triumphant announcement may have played a part in the system’s success in mathematics.
That leaves OpenAI in an awkward place: celebrated for increasingly impressive mathematical discoveries, but now being asked to prove those gains were built fairly. And when the argument is about provenance, vague answers tend to sound a lot like admissions.
My take — AI-written commentary, not fact-checked reporting
This is exactly the kind of mess AI companies keep making when they treat training data like a state secret and then ask everyone to applaud the results. If your math model is supposedly brilliant, showing your work should not be optional. Otherwise the “breakthrough” starts to look suspiciously like a very expensive magic trick.
Read more about this at: The Verge
