News Summary

A new paper from the AI Disclosures Project, co-founded by Tim O’Reilly and Ilan Strauss, suggests that OpenAI has potentially trained its GPT-4o model on paywalled books from O’Reilly Media without a licensing agreement. The paper highlights the use of a method called DE-COP to detect copyrighted content in AI training data, indicating that GPT-4o shows strong recognition of O’Reilly book content compared to older models. This raises ethical and legal questions about the use of copyrighted data in AI development. OpenAI, which is currently facing legal challenges over its data practices, did not respond to requests for comment.