OpenAI Battles Court Order to Hand Over 20 Million ChatGPT Conversations in Copyright Case

OpenAI has asked a U.S. federal court in New York to overturn an order that requires it to turn over about 20 million anonymised chat logs from its ChatGPT users, in a major legal showdown with The New York Times and other news outlets.

The logs are sought as evidence in the plaintiffs’ lawsuit accusing OpenAI of using their copyrighted content without permission to train its AI models. The plaintiffs argue the user chats are necessary to determine how often and in what way ChatGPT reproduces their content.

OpenAI counters that the requirement is overly broad and threatens user privacy, noting that “99.99 %” of the requested transcripts have no relevance to the underlying copyright claims. The company warned that the order “forces us to turn over tens of millions of highly personal conversations from people who have no connection to the Times’ … lawsuit.”

The court order was issued by Magistrate Judge Ona Wang, who asserted that existing privacy protections (including de-identification and protective orders) were adequate and that the motion for such discovery was warranted.

The case raises far-reaching implications about how AI companies handle user data, what limits exist for discovery in copyright disputes, and where the balance lies between protecting privacy and enforcing intellectual property law. OpenAI’s appeal will test whether courts can force mass disclosure of user conversations in AI-training disputes.