OpenAI Got Caught Lying to a Court — and Deleting Billions of ChatGPT Logs

A group of publishers led by The New York Times, including the New York Daily News, has filed a motion in federal court in the Southern District of New York asking the judge to sanction OpenAI for alleged misconduct in their ongoing copyright dispute — accusing the company of lying to the court about its technical ability to search its training data and ChatGPT logs for publishers' content.

What the allegations say

According to the plaintiffs, OpenAI initially told the court it could not search its training datasets or ChatGPT output logs for the publishers' copyrighted content — even though the company had, in fact, already run such searches before being sued. The publishers also allege OpenAI deleted billions of relevant ChatGPT conversations or made them unsearchable.

What the deposition revealed

The discrepancies surfaced through testimony from OpenAI's Vinnie Monaco, who leads privacy engineering at the company. According to the publishers, a follow-up deposition of Monaco revealed that OpenAI already had tools to search training data, build searchable de-identified datasets of ChatGPT conversations, and search for news publishers' content. The allegations are laid out in a 52-page filing submitted to the court.

What this means in practice

The publishers are asking the court for sanctions, including attorneys' fees, along with a formal finding that ChatGPT's conversation logs show misuse of their copyrighted material. If granted, the motion could set a significant precedent — not just for the underlying copyright dispute, but for how honestly AI companies are required to disclose their own technical capabilities to courts.

This material is for informational purposes only.