Mathematician Andreas Thom accused OpenAI of dishonesty and a lack of transparency about whether his prior ChatGPT interactions entered its training data, following the company's non-sofic groups result that built on work by Thom and Gábor Kun. Thom's concerns echo those of New York University professor Tristan Buckmaster, who questioned whether OpenAI's Codex benefited from his unpublished work. Thom says only OpenAI can prove whether user data was used, calling undisclosed use ethically indefensible.
No score is assigned. Sources and their independence are shown in the citation chain below.