A coalition of media organizations, including The New York Times, The New York Daily News, and 15 other publishers, petitioned a federal court on Thursday to impose sanctions against OpenAI, accusing the artificial intelligence company of withholding crucial evidence in ongoing copyright infringement lawsuits. The motion alleges that OpenAI has failed to fully disclose information pertinent to how its AI models are trained and utilized, thereby obstructing fact discovery and violating court rules.

The legal disputes stem from a lawsuit filed by The New York Times in late 2023, which claims OpenAI unlawfully used the newspaper’s copyrighted content to train ChatGPT and other AI tools without permission. Following The Times’ suit, multiple other publishers initiated similar litigation against OpenAI. In 2023, these cases were consolidated by a judicial panel to streamline proceedings.

In their sanctions request, the media organizations contend that OpenAI has not provided access to internal data sets and output logs that would clarify the company’s training processes. The filing argues that such evidence is essential to evaluating OpenAI’s defense, which is based on a claim of fair use. Instead, the publishers assert, OpenAI chose to impede the discovery process, prolonging the litigation and complicating substantive review of the claims.

OpenAI responded to the filing through spokesperson Drew Pariser, stating that the companies bringing suit are resorting to “blatantly false allegations” and invasive demands unrelated to the case. Pariser emphasized OpenAI’s commitment to user privacy and maintained that The Times’ legal claims have weakened over time, forcing the media outlet to drop certain allegations. The company continues to deny any wrongdoing, asserting that it respects content creators’ rights and that ChatGPT is not a replacement for subscriptions to The Times.

Legal experts note that motions seeking sanctions are relatively uncommon and require judges to actively intervene in disputes over compliance with discovery obligations. Robin Feldman, a law professor at the University of California, San Francisco, described such motions as unusual and indicative of contentious litigation dynamics.

The Times was the first major American publisher to pursue legal action against OpenAI over alleged copyright violations tied to text-based content. The lawsuit also named Microsoft, a key partner of OpenAI, although Microsoft has strongly denied similar allegations and is not targeted in the current sanctions filing. The broader consolidated litigation includes cases from authors such as Sarah Silverman, John Grisham, Jonathan Franzen, and George R.R. Martin.

OpenAI and other AI developers rely on extensive data sets, which may include copyrighted material, to train machine learning models. The companies argue their use falls under fair use doctrines, given that the data is transformed and repurposed for new applications without substituting for the original works.

Following the deposition of an OpenAI employee, which the publishers say revealed ongoing misrepresentations by OpenAI about its ability to locate specific content in its training data and logs, the media entities accuse the company of bad faith conduct. Steven Lieberman, attorney for The Daily News and other plaintiffs, said OpenAI has repeatedly misled courts and litigants. Lead counsel Ian C. Brosky of The Times’ legal team echoed this view, asserting that OpenAI has consistently lied to various stakeholders.

The publishers are seeking financial penalties and other sanctions intended to compel OpenAI’s compliance. Their request does not seek similar measures against Microsoft. The dispute highlights ongoing legal challenges at the intersection of artificial intelligence development and copyright law.