OpenAI Fights to Limit NYT’s Access to 120 Million ChatGPT Logs Amid User Privacy Fears

OpenAI and the New York Times, in a joint court filing, have urged a federal judge to establish a confidential settlement session for August 7.  Despite the phrase “settlement,” this isn’t about ending the litigation. Instead, the session will focus on whether NYT lawyers can obtain access to 120 million user conversations - an amount OpenAI is calling unreasonable.

“The New York Times and other plaintiffs have made a sweeping and unnecessary demand in their baseless lawsuit against us: retain consumer ChatGPT and API customer data indefinitely. This fundamentally conflicts with the privacy commitments we have made to our users. It abandons long-standing privacy norms and weakens privacy protections,” said OpenAI COO Brad Lightcap. 

Calling the lawsuit “baseless,” OpenAI claims that it is an “overreach” and risks user privacy without helping resolve the lawsuit. 

The NYT is trying to find proof that OpenAI’s chatbot copied and republished copyrighted news articles without permission.  That kind of evidence could seriously hurt OpenAI’s case.  But to get it, they want to search through a massive batch of ChatGPT logs, including chats OpenAI had previously promised were deleted.  OpenAI is pushing back. The company thinks that a far smaller sample of roughly 20 million logs would be enough to investigate how often ChatGPT might regurgitate news material.  But the NYT rejected the offer. OpenAI is now warning the court that going along with the complete request could delay the lawsuit for months and raise severe user privacy issues. 

If the NYT gets its way, OpenAI may be forced to restore deleted chats and keep them stored for longer - something ChatGPT users never expected.  The legal battle is unfolding at a time when AI privacy is already under scrutiny.  Some users fear that their personal, even sensitive, conversations with ChatGPT could be exposed in court.  Meanwhile, Microsoft, OpenAI’s partner and co-defendant, is fighting its own battle to stop the NYT from looking into internal tools like its ChatGPT-powered assistant. As the clock ticks down to the conference on August 7, the stakes are higher than just two companies.  The court’s decision could set a major precedent for how far copyright cases can go and how private your AI chats really are.

Read More: New Instagram Feature Lets Creators See Exactly When Viewers Like