OpenAI accuses the New York Times of invading user privacy in discovery dispute

OpenAI published a blog post publicly accusing the New York Times of seeking to “invade” its users’ privacy, escalating a discovery dispute inside the copyright lawsuit the Times and other news organisations had brought against OpenAI and Microsoft since December 2023. The dispute concerned a magistrate judge’s order that OpenAI produce 20 million anonymised ChatGPT conversation logs, spanning December 2022 to November 2024, as a sample from which plaintiffs could search for evidence that the model reproduced their copyrighted articles. OpenAI had proposed instead to run keyword searches and hand over only conversations touching the plaintiffs’ specific works — filtering the magistrate judge rejected as insufficient — and argued that a full production would expose sensitive personal material from tens of millions of users who had no connection to the case.

The Times disputed OpenAI’s framing, saying the order applied only to anonymised logs covered by a protective order restricting who could view them and for what purpose, and that no user’s identity was at risk. OpenAI filed a motion the same day seeking to reverse the order.

The exchange showed both parties litigating in public as well as in court, each casting the other as the party threatening user interests, in a discovery fight that continued into early 2026 before a district judge affirmed the production order.