A sanctions motion in the Southern District of New York says OpenAI spent two years claiming it couldn't search ChatGPT logs for copyrighted news content — then a deposition showed it already had.
Discovery fights in AI copyright cases usually turn on scale: petabytes of scraped text, models too complex to audit, logs too vast to search. OpenAI has leaned on that framing for more than two years in its defense against the New York Times and other publishers. On Thursday, the Times and its co-plaintiffs told a federal judge that framing was a lie.
In a 52-page motion for sanctions filed in the U.S. District Court for the Southern District of New York, the Times, the Daily News, the Center for Investigative Reporting, the Intercept and CNET parent Ziff Davis accused OpenAI of a two-year effort to conceal its ability to search training datasets and output logs for copies of their journalism, according to Variety. Bloomberg Law characterized the filing as alleging a “deliberate and systemic effort to obstruct discovery.” Ars Technica‘s headline on the filing went further, saying OpenAI “faked inability to search training data, hid billions of logs.”
the monaco deposition
By the plaintiffs’ account, the concealment only surfaced after a second deposition of Vinnie Monaco, who leads privacy engineering at OpenAI. In that February session, per Variety, Monaco revealed the company had in fact searched training datasets and output data — despite OpenAI’s earlier claims that such searches were not feasible. The outlets also allege OpenAI deleted logs in violation of the court’s preservation orders, a claim that, if borne out, would compound the discovery dispute into a spoliation problem.
“Instead of just producing that evidence at the start of the case and focusing on the merits of its fair use defense, OpenAI chose obstruction,” the news organizations wrote through their attorneys, according to Variety. The motion asks the court to formally find that OpenAI made false claims and destroyed output logs, to award the plaintiffs’ attorneys’ fees, and to impose whatever other remedy it deems fit. “Each of the remedies requested is an appropriate exercise of this Court’s inherent authority and necessary both in response to OpenAI’s misconduct and as a deterrent to those who might contemplate following their example,” the filing states.
dueling statements
Rian Crosby, the Times’ lead attorney and a partner at Susman Godfrey, was blunt in a statement reported by Variety: OpenAI “lied to The Times, The Daily News Plaintiffs, the public, and the court.” Crosby added that OpenAI “claimed searching ChatGPT outputs for copies of The Times’ and the Daily News Plaintiffs’ content was infeasible, burdensome, and invasive of users’ privacy – while at the same time concealing that it had already done such searches,” and argued the concealment undercuts OpenAI’s substantive defense: “If OpenAI genuinely believed that copying our clients’ journalism was fair and legal, it wouldn’t have hid the truth about having done it.”
OpenAI disputes the characterization entirely. A company spokesperson told Variety the Times was “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,” adding that OpenAI would “continue defending our users’ privacy and the long-established principles of fair use.” That leaves the court to sort out a sharply contested factual record — one side describing a data-privacy defense, the other describing a cover story — before it can even reach the underlying fair-use question.
the wider docket
The dispute sits inside the oldest and largest of the news-industry copyright suits against OpenAI. The Times sued OpenAI and Microsoft in 2023, becoming the first news outlet to bring such a case; the Daily News, the Center for Investigative Reporting and Ziff Davis filed separate suits in 2024 alleging similar conduct. The Times has also sued Perplexity over comparable allegations, even as it pursues licensing deals with other AI players — it struck a content-licensing agreement with Amazon for “AI-related uses” last year while building its own AI tools internally, per Variety. That dual-track strategy — litigate against some, license with others — has defined how publishers are approaching the generative-AI economy, and it isn’t cheap: the Times has spent more than $28 million on litigation costs against AI companies, including $4.2 million in the first quarter of 2026 alone, according to Variety’s reporting.
What happens next will hinge on how the SDNY court weighs the sanctions motion — whether it credits the plaintiffs’ spoliation and concealment claims, and what remedy, if any, it imposes. For data companies training on scraped or licensed content, the case is a reminder that discovery obligations in these suits are increasingly as consequential as the fair-use merits themselves: claims about what can and can’t be searched inside a training pipeline are now themselves evidence, subject to depositions, preservation orders, and — if this motion succeeds — sanctions.
It claimed searching ChatGPT outputs for copies of The Times' and the Daily News Plaintiffs' content was infeasible, burdensome, and invasive of users' privacy – while at the same time concealing that it had already done such searches. If OpenAI genuinely believed that copying our clients' journalism was fair and legal, it wouldn't have hid the truth about having done it.
