Mathematicians want proof OpenAI didn't use their work (www.theverge.com)

Mathematicians are raising serious concerns about OpenAI's ethical practices regarding the training data behind its highly publicized mathematical breakthroughs. Following a controversy about whether OpenAI leveraged unpublished research from mathematicians, Andreas Thom has publicly accused the company of dishonesty and lack of transparency. His suspicions were fueled by the fact that OpenAI's recent announcement contained results closely related to his work on non-sofic groups, a field he has been actively researching. Despite OpenAI’s claims to have not directly used specific user data, Thom argues that their ambiguous statements raise questions about the potential indirect influence of private interactions on their models. The significance of this debate extends beyond individual grievances; it raises crucial ethical implications for the AI/ML community. Researchers like Thom are not able to audit OpenAI’s training processes, leaving them at a disadvantage in understanding if their contributions inform AI advancements. This lack of transparency could deter collaboration and encourage secrecy among mathematicians, fearing that their ideas might be appropriated without credit in a competitive landscape dominated by large tech firms. As OpenAI faces growing scrutiny, this incident prompts larger discussions about accountability, trust, and the future of collaborative research in mathematics and AI.