More questions about whether researchers can trust OpenAI with unpublished math

pred_ · hn · 2026-09-10

Mathematicians across Mastodon, X, and Bluesky are raising fresh concerns over whether researchers can trust OpenAI with unpublished mathematical work. The debate centers on the risks of feeding unpublished proofs and research into OpenAI models: potential leakage, scooping, or use in training. The discussion continues to unfold on Hacker News, touching on the boundary between academic confidentiality and frontier AI companies.

Original post →

More from Companies & People

Companies & People channel →