News Publishers Ask Judge to Sanction OpenAI, Alleging It Hid Evidence

A New York Times-led coalition says OpenAI claimed it could not search its training data for copyrighted journalism — while an engineer's deposition showed it had already done exactly that, before the first lawsuit was filed.

Jul 10, 2026Source: TechCrunch

News Publishers Ask Judge to Sanction OpenAI, Alleging It Hid Evidence

Key Points

1.The New York Times, the New York Daily News and more than a dozen other news organizations — among them the Chicago Tribune, the Santa Rosa Press Democrat, Ziff Davis, The Intercept and the Center for Investigative Reporting — moved for sanctions against OpenAI on July 9 in the Southern District of New York, alleging discovery misconduct.
2.The motion leans on the April deposition of OpenAI privacy engineering lead Vinnie Monaco, who plaintiffs say confirmed the company had searched its training corpus and outputs for copyrighted journalism, and had assembled a database of roughly 78 million de-identified ChatGPT conversations starting before the first suit was filed.
3.Plaintiffs also allege OpenAI deleted or made unsearchable billions of chat logs — some after a preservation order — and ask the court to reject the 20-million-log sample OpenAI produced; OpenAI says the claims are 'blatantly false' and that the Times is invading users' privacy.

The copyright fight over ChatGPT has turned into a fight about what OpenAI told the court. On Thursday, July 9, a coalition of six news organizations asked a federal judge in Manhattan to sanction OpenAI for what they call discovery misconduct: claiming it was technically unable to search its training data and chat logs for copyrighted journalism, while internally having already done so.

Who filed

The motion, filed in the Southern District of New York, comes from The New York Times, the New York Daily News and more than a dozen other news organizations, among them the Chicago Tribune and other Tribune Publishing and MediaNews Group papers, the Santa Rosa Press Democrat, Ziff Davis, The Intercept and the Center for Investigative Reporting. The underlying case, filed by the Times on December 27, 2023 against OpenAI and co-defendant Microsoft, is before Judge Sidney H. Stein and survived a motion to dismiss in April 2025.

The deposition

The pivot is Vinnie Monaco, OpenAI's privacy engineering lead, deposed as a corporate witness in April after a magistrate judge ordered him back for three and a half additional hours, finding him unprepared for the first session. Plaintiffs say Monaco confirmed OpenAI had run internal searches of its training corpus and model outputs specifically for copyrighted journalism, and that the company had built a database of roughly 78 million de-identified ChatGPT conversations — work that began before the first news plaintiff sued. Separately, they describe a regurgitation-detection tool, a "Bloom" filter built under a project called Giraffe, created shortly after the suit landed.

What they want

Beyond sanctions, the publishers ask the court to reject the 20-million-chat-log sample OpenAI produced as an unreliable stand-in for the full record. Coverage of the motion describes the relief sought as monetary penalties, special jury instructions and other measures. They allege OpenAI deleted or rendered unsearchable billions of relevant conversations, some after a court order to preserve them. "If OpenAI genuinely believed that copying our clients' journalism was fair and legal, it wouldn't have hid the truth about having done it," said plaintiffs' counsel Ian B. Crosby. Co-counsel Steven Lieberman put it more bluntly: the motion "asks the court to punish OpenAI for hiding and destroying evidence showing how ChatGPT was trained on stolen journalism."

OpenAI's answer

Spokesperson Drew Pusateri said: "As the Times' case weakens and they've been forced to drop claims against us, they're persisting with their efforts to invade the privacy of people who have nothing to do with this case, including by making these blatantly false allegations." Judge Stein has not ruled.

Why It Matters

Fair use has been OpenAI's whole defense, and it rests on the claim that ChatGPT rarely regurgitates its training data. If the court accepts that OpenAI could measure regurgitation all along — and destroyed the logs that would show it — the argument shifts from a legal question about copying to an evidentiary one about concealment, where judges have far blunter instruments and far less patience.

Filed underOpenAI AI Copyright #New York Times #fair use #litigation