Key facts
- The New York Times and other newspapers are seeking sanctions against OpenAI.
- The lawsuit alleges OpenAI lied to the court about its ability to search its AI models for copyrighted articles.
- Newspapers claim OpenAI conducted searches of copyrighted material prior to the lawsuit and deleted relevant data.
- The motion requests attorneys' fees and a court finding of misuse of copyrighted works.
A coalition of newspapers, spearheaded by The New York Times and the New York Daily News, has formally requested sanctions against OpenAI in their ongoing copyright infringement lawsuit. The newspapers allege in a filing with a federal court in Manhattan that OpenAI deliberately misled the court regarding its capabilities to search its large language models for copyrighted material.
According to the newspapers' filing on Thursday, OpenAI falsely claimed it could not search its systems for their articles, while simultaneously hiding the fact that such searches had already been conducted, even before the initial lawsuit was filed. The plaintiffs also contend that OpenAI has deleted or rendered unsearchable billions of ChatGPT conversations that are relevant to the case.
The legal action seeks sanctions against OpenAI, including reimbursement for attorneys' fees, and a judicial determination that the company's internal chat logs demonstrate the misuse of their copyrighted works. OpenAI has not yet responded to a request for comment on the motion.
The original lawsuit, initiated by The Times in 2023, accuses OpenAI and its major investor Microsoft of utilizing millions of articles without authorization to train the AI model powering ChatGPT. This case is part of a broader trend of copyright holders, including authors, artists, and music labels, taking legal action against technology firms like OpenAI, Anthropic, and Meta Platforms for alleged unauthorized use of their content in AI system training.
Ian Crosby, lead attorney for The New York Times, stated that OpenAI had deceived the plaintiffs, the public, and the court for over two years by claiming that searching ChatGPT outputs for copyrighted content was infeasible, burdensome, and an invasion of user privacy, all while concealing its own prior searches.