Mathematicians Accuse OpenAI of Using Their Work Without Proper Acknowledgment
A mathematician, Andreas Thom, has accused OpenAI of unethical and dishonest behavior after the company announced a breakthrough in non-sofic groups. Thom raised concerns that his prior interactions with ChatGPT may have contributed to this success without proper acknowledgment or transparency from OpenAI. He noted that one of the results highlighted by OpenAI built heavily on previous work he did alongside Gábor Kun, and while the company later amended its writeup, it failed to provide sufficient evidence proving their training data was not influenced by user inputs. Thom emphasized his inability to reverse-engineer OpenAI’s models or determine whether his contributions were used. He called for greater transparency from the AI firm in disclosing datasets and clarifying how they use external input. This issue echoes a similar controversy involving Tristan Buckmaster, who questioned if OpenAI had benefited indirectly from user data during its recent Millennium Prize breakthrough. The broader concern is that researchers are being left without clear answers or accountability as AI companies like OpenAI make increasingly complex mathematical discoveries using opaque methods.
Topics: academic integrity AI data transparency mathematics research OpenAI