OpenAI has told a federal judge that 5% of chat logs were deleted, not hidden, in the NY Times copyright case. The company argues the publishers lack evidence of systematic regurgitation of their articles.
TL;DR
- OpenAI claims it deleted 5% of chat logs, not hidden them, in the NY Times copyright case.
- The company argues the publishers lack evidence of systematic regurgitation of their articles.
- Publishers seek sanctions, arguing OpenAI's actions have hampered their case.
What happened
OpenAI Inc. has informed a federal judge that news organizations, including The New York Times Co. and The Daily News, are seeking sanctions against it for alleged discovery misconduct in a copyright lawsuit. The AI company claims it has followed court orders to preserve user chatbot conversations and has turned over its training data and tens of millions of ChatGPT outputs. According to OpenAI, it has gone to 'herculean efforts' to comply with the court's directives.
OpenAI argues that the publishers' accusations of hiding its ability to search training data and output logs for their articles are an attempt to 'rewrite history' and avoid proving their case. The company claims that the court's preservation order did not require 'wholesale' preservation or for OpenAI to stop all deletions, allowing users to delete chats that were more than 30 days old as of the date of the court's directive. This resulted in about 5% of the logs the publishers requested being deleted, but OpenAI maintains they were replaced with other logs that didn't bias the sample.
The publishers argue that OpenAI's actions have prevented them from establishing the foundation of their case—the extent to which OpenAI scrapes, trains, and copies its content to respond to user prompts. They seek to prohibit OpenAI from relying on the produced output logs and are asking for attorneys' fees associated with the alleged obstruction. Microsoft, a co-defendant, also opposed the publishers' bid, arguing the requested restrictions against OpenAI would unduly prejudice Microsoft.
Why it matters
This case has significant implications for AI developers and startups, as it could set a precedent for how AI companies handle data scraping and copyright issues. The outcome may influence how AI models are trained and how companies respond to similar legal challenges in the future. For investors, the case highlights the legal risks associated with AI technologies and the importance of compliance with data preservation orders.
The competitive angle is also crucial, as the case could impact the relationship between AI companies and publishers. If the publishers succeed in their sanctions bid, it could hamper OpenAI's ability to mount a defense and potentially lead to a favorable outcome for the publishers. This could set a precedent for future cases involving AI and copyright infringement. However, the case also raises questions about the technical capabilities of AI companies and their ability to comply with court orders.
Key facts
- OpenAI claims it deleted 5% of chat logs, not hidden them, in the NY Times copyright case.
- The company argues the publishers lack evidence of systematic regurgitation of their articles.
- Publishers seek sanctions, arguing OpenAI's actions have hampered their case.
- OpenAI claims it has followed court orders to preserve user chatbot conversations and has turned over its training data and tens of millions of ChatGPT outputs.
- The case is In Re: OpenAI Inc. Copyright Infringement Litig., S.D.N.Y., No. 25-md-03143, opposition filed 8/14/26.
- OpenAI is represented by Williams & Connolly LLP, Keker Van Nest & Peters LLP, Latham & Watkins LLP, and Morrison & Foerster LLP.
- Microsoft is represented by Faegre Drinker Biddle & Reath LLP and Orrick, Herrington & Sutcliffe LLP.
- The publishers are represented by Susman Godfrey LLP, Rothwell Figg Ernst & Manbeck PC, Loevy & Loevy, Klaris Law PLLC, and Ruttenberg IP Law.
Context
This case is part of a broader conversation about the intersection of AI and copyright law. As AI technologies continue to advance, questions about data scraping, training, and copyright infringement are becoming increasingly important. The outcome of this case could have significant implications for AI developers, startups, and investors, as well as for publishers and other content creators. It highlights the need for clear guidelines and regulations in this rapidly evolving field.
