Editing Openai/69110721-f258-8001-9e77-11494b128269 (section)

===== 1. He asserted that OpenAI used large volumes of copyrighted data without adequate permission in its training of models like ChatGPT. pbs.org<ref>{{cite web|title=pbs.org|url=https://www.japantimes.co.jp/business/2024/10/24/tech/openai-researcher-copyright-law/|publisher=japantimes.co.jp|date=2024-10-24|access-date=2025-11-10}}</ref> - He specifically argued that the “fair use” defence (in U.S. copyright law) may not apply cleanly to how these models are trained. VKTR.com<ref>{{cite web|title=VKTR.com|url=https://www.vktr.com/ai-news/former-openai-researcher-stirs-fair-use-debate/|publisher=vktr.com|access-date=2025-11-10}}</ref> - He claimed that because models trained on data from creators (authors, programmers, etc) could produce substitutes, it harms those creators and the ecosystem. The Times of India<ref>{{cite web|title=The Times of India|url=https://timesofindia.indiatimes.com/world/us/suchir-balaji-openai-chatgpt-whistleblower/articleshow/116312596.cms|publisher=The Times of India|access-date=2025-11-10}}</ref> =====
# He expressed concern that OpenAI’s business model (in his view) was not simply a research nonprofit pursuit, but something that could compete with the very creators whose data was used. For example: “It doesn’t feel right to be training on people’s data and then competing with them in the marketplace.” AP News<ref>{{cite web|title=AP News|url=https://apnews.com/article/283e70b31d34ebb71b62e73aafb56a7d|publisher=AP News|access-date=2025-11-10}}</ref>