Report / Tech
Published by NOISE
Mathematicians demand proof OpenAI did not use their work
From Morning NOISE.
The Verge reports mathematician Andreas Thom accused OpenAI of dishonesty over its non-sofic groups result, saying the company failed to prove his prior ChatGPT interactions were excluded from training data. The dispute follows Tristan Buckmaster questioning whether his Codex use aided OpenAI's Navier-Stokes solution, which OpenAI said did not use specific user data while adding it could not rule out help from de-identified usage data. Thom argues only OpenAI can verify its pipeline and should disclose datasets, underscoring uncertainty about what training data major AI models actually used.
Reporting used
Saved versions
These dates record changes to NOISE’s copy. They are not event dates.