New York Times accuses OpenAI of hiding evidence in copyright case

The New York Times Company, Daily News LP, and other media publishers escalated their copyright lawsuit against OpenAI by filing a sanctions motion in federal court, alleging the company deliberately withheld evidence and made false representations to the court about its technical capabilities TechCrunch.

The motion, reported on July 9, 2026, centers on OpenAI's alleged obstruction of discovery in the ongoing copyright litigation. According to the publishers' counsel, "The evidence is in OpenAI's training data sets and ChatGPT output logs." Steven Lieberman, representing the Daily News, alleged in the filing that "For two years, OpenAI has been making misrepresentations to the court regarding its ability to search for Daily News content in its training data sets and output logs." Ars Technica

Ian B. Crosby, counsel for the plaintiffs, made a stronger assertion, stating that "OpenAI lied to the Times, the Daily News plaintiffs, the public and the court." The publishers claim OpenAI chose obstruction over compliance, despite having the technical means to locate their content within ChatGPT's training and operational records World IP Review.

Requested Remedies

The publishers are asking the court to impose discipline and sanctions on OpenAI. Their specific requests include barring OpenAI from using a 20 million chat log sample as evidence in the case, treating ChatGPT logs as demonstrating substantial regurgitation and grounding of the plaintiffs' copyrighted content, and requiring OpenAI to pay the publishers' legal fees incurred in the discovery dispute.

OpenAI's Response

OpenAI's spokesperson Drew Pusateri rejected the allegations, countering that the Times is attempting to gain access to private user conversations as its legal position weakens. The company has not publicly released detailed technical responses to the specific claims about dataset searchability and log retention capabilities TechCrunch.

Context

This sanction motion represents a significant escalation in the copyright dispute, which centers on whether OpenAI's training of ChatGPT on publishers' articles without permission constitutes infringement. The case reflects broader tension in the AI industry over training data sourcing, with discovery disputes becoming a focal point. The publishers' emphasis on the existence and accessibility of the evidence—rather than its absence—suggests they believe OpenAI possesses the capability to comply but has chosen not to.