Key Takeaways:
The New York Times and The Daily News accused OpenAI of concealing its ability to search customer chat logs and training datasets for copyrighted material, escalating a two-year-old lawsuit over the use of journalism to train generative AI models.
"OpenAI has repeatedly represented that it cannot search its training data or user chat logs for specific copyrighted works, but internal evidence suggests otherwise," the newspapers said in a filing in the Southern District of New York. The case, consolidated before Judge Sidney Stein, is one of the most closely watched in a wave of more than 125 copyright lawsuits filed against AI companies in the US since 2023.
The newspapers allege that OpenAI trained its GPT models on millions of articles without permission and that ChatGPT reproduced their journalism verbatim in user outputs. The accusation of hidden evidence could carry significant consequences — if the court finds OpenAI failed to preserve or produce relevant data, it could face sanctions including adverse inference instructions that would allow a jury to assume the missing evidence would have harmed OpenAI's case.
The dispute centers on whether OpenAI's systems can identify and remove copyrighted content from both its training datasets and the responses ChatGPT generates. The Times has sought access to OpenAI's training data and output logs to prove infringement, while OpenAI has argued that searching such data at scale is technically infeasible. The newspapers' latest filing claims that OpenAI's own internal tools contradict that position.
Discovery has intensified in recent months. In January, the court ordered OpenAI to produce 20 million output logs, and in March, an additional 78 million and 10 million log reservoirs. The Times and Daily News are part of the "Newspapers Action" within the broader In Re OpenAI Copyright Infringement multidistrict litigation, which also includes claims from The Center for Investigative Reporting, Ziff Davis, and other publishers. A trial date has not yet been set.
The outcome could reshape the economics of generative AI. OpenAI, valued at more than $300 billion in its latest funding round, relies on vast datasets of copyrighted material to train its models. If courts rule that such training requires licensing, AI companies could face billions of dollars in annual royalty costs — a risk that has already pushed some firms, including Disney and OpenAI, to sign formal licensing deals. Microsoft, OpenAI's largest backer, is also named as a defendant in the case.
This article is for informational purposes only and does not constitute investment advice.
