The court ordered OpenAI in its case against The New York Times to produce 20 million chat logs. OpenAI slams the court order, but it’s what they offered previously.

According to ArsTechnica (OpenAI slams court order that lets NYT read 20 million complete user chats, written by Jon Brodkin and available here), OpenAI wants a court to reverse a ruling forcing the ChatGPT maker to give 20 million user chats to The New York Times and other news plaintiffs that sued it over alleged copyright infringement. Although OpenAI previously offered a “sample size” of 20 million user chats as a counter to the NYT’s demand for 120 million, the AI company says a court order requiring production of the chats is too broad.

“The logs at issue here are complete conversations: each log in the 20 million sample represents a complete exchange of multiple prompt-output pairs between a user and ChatGPT,” OpenAI said Wednesday in a filing in US District Court for the Southern District of New York. “Disclosure of those logs is thus much more likely to expose private information [than individual prompt-output pairs], in the same way that eavesdropping on an entire conversation reveals more private information than a 5-second conversation fragment.”

OpenAI’s filing said that “more than 99.99%” of the chats “have nothing to do with this case.” It asked the district court to “vacate the order and order News Plaintiffs to respond to OpenAI’s proposal for identifying relevant logs.” OpenAI could also seek review in a federal court of appeals.

In her order last Friday, New York Magistrate Judge Ona T. Wang stated: “Whether or not the parties had reached agreement to produce the 20 million Consumer ChatGPT Logs in whole—which the parties vehemently dispute—such production here is appropriate. OpenAI has failed to explain how its consumers’ privacy rights are not adequately protected by: (1) the existing protective order in this multidistrict litigation or (2) OpenAI’s exhaustive de-identification of all of the 20 million Consumer ChatGPT Logs.” She added in a footnote: “As News Plaintiffs point out, OpenAI ‘has spent the last two and a half months processing and deidentifying this 20 million record sample.’”

Judge Wang also stated: “Further, OpenAI has consistently relied on the use of Cochran’s Formula by Judge Keulen in Concord Music Group, Inc. v. Anthropic PBC, 24-CV-3811 (EKL) (SVK), 2025 WL 1482734 (N.D. Cal. May 23, 2025), in support of its previous proposed methodology for conversation data sampling,…but fails to explain why Judge Keulen’s subsequent order directing production of the entire 5-million record sample to the plaintiff in that case is not similarly instructive here.” So, she ordered OpenAI to produce the logs by November 14 (i.e., today) or “within 7 days of completing the de-identification process”.

In response to that point, OpenAI stated: “News Plaintiffs did not even reference Concord in their motion…OpenAI never had a chance to explain that unlike OpenAI, Anthropic had affirmatively proposed wholesale production of 5 million user logs without any apparent concern for the privacy implications. The Concord court thus never had to “direct[]” wholesale production.”

Am I missing something? As referenced above, in this filing back in August that detailed its process for generating a sample of the logs, OpenAI stated that it “has in fact proffered a 20 million conversation sample”, arguing: “This size is surely more than enough to conduct appropriate analyses relevant to Plaintiffs’ claims.” The issue now seems to be whether OpenAI agreed to produce the chat logs in whole. But, as Judge Wang stated, there’s a protection order in place and OpenAI has spent 2 1/2 months “deidentifying” the 20 million record sample.

OpenAI slams the court order, but it’s what they offered previously. Sort of. It will be interesting to see if they take the ruling to appeal.

So, what do you think? Do you think that OpenAI has a valid argument? Please share any comments you might have or if you’d like to know more about a particular topic.

Disclaimer: The views represented herein are exclusively the views of the author, and do not necessarily represent the views held by my employer, my partners or my clients. eDiscovery Today is made available solely for educational purposes to provide general information about general eDiscovery principles and not to provide specific legal advice applicable to any particular circumstance. eDiscovery Today should not be used as a substitute for competent legal advice from a lawyer you have retained and who has agreed to represent you.