OpenAI is escalating its public defense against The New York Times, publishing a forceful statement last Wednesday arguing that the newspaper’s demand for 20 million ChatGPT user conversations represents an unprecedented violation of user privacy. The public stance comes even as a federal judge has already ruled that OpenAI must turn over the anonymized logs as part of the Times’ ongoing copyright lawsuit against the company and its partner Microsoft, according to the Business Insider.

The high-stakes battle now sits at the intersection of journalism, copyright law, and the privacy expectations of hundreds of millions of AI users — raising new questions about how much personal data may enter the legal system as generative AI becomes embedded in daily life.

In a November 7 order, Magistrate Judge Ona Wang determined that it was “appropriate” for OpenAI to produce the requested logs. In her reasoning, Wang wrote that OpenAI had not shown how its users’ privacy would be compromised, given the protections already in place:

In short, the judge concluded that OpenAI’s concerns were insufficient and that privacy safeguards were already robust.

But in its public statement, authored by Dane Stuckey, the company’s Chief Information Security Officer, OpenAI countered that the ruling threatens user trust, arguing:

“This demand disregards long-standing privacy protections, breaks with common-sense security practices, and would force us to turn over tens of millions of highly personal conversations from people who have no connection to the Times’ baseless lawsuit.”

The company emphasized that 800 million people use ChatGPT weekly — many for deeply personal matters — and that users entrust the platform with “files, credentials, memories, searches, payment information, and AI agents that act on their behalf. We treat this data as among the most sensitive information in your digital life — and we’re building our privacy and security protections to match that responsibility."

The New York Times filed its copyright infringement lawsuit in 2023, alleging that OpenAI and Microsoft trained AI models on Times journalism without permission, allowing ChatGPT to replicate or paraphrase its reporting.

To understand how users interact with ChatGPT — and whether it returns Times content — the newspaper is seeking a random sample of 20 million chats.

The Times argues that:

A spokesperson for the Times sharply criticized OpenAI’s statement:

“No ChatGPT user's privacy is at risk. The court ordered OpenAI to provide a sample of chats, anonymized by OpenAI itself, under a legal protective order. This fear-mongering is all the more dishonest given that OpenAI’s own terms of service permit the company to train its models on users’ chats and turn over chats for litigation.” OpenAI says it supports independent journalism and continues to work with a wide range of publishers and news organizations.

The Times also noted that OpenAI already collects and uses chat data — a point it argues weakens OpenAI’s privacy claims.

OpenAI says this is not merely a discovery dispute — it’s a matter of protecting user trust at scale.

According to the company:

OpenAI also challenged the Times’ justification that another AI company handed over 5 million chats in an unrelated case — arguing the situations are not comparable.

The company says it offered narrower, privacy-preserving alternatives, including:

The Times rejected those proposals, arguing that narrower alternatives would not provide the transparency needed to determine whether ChatGPT reproduces or relies on Times reporting.

According to the ruling, the data OpenAI must produce includes the following elements:

The court order does not cover any data from:

This means only consumer ChatGPT conversations are part of the sample, and all business-tier and organizational data remains outside the scope of the request.

Under the terms of the court order, the anonymized and “scrubbed” chats can only be examined by the following authorized parties and under strict access conditions:

OpenAI emphasizes that it will oppose any attempt to make the anonymized data public, arguing that safeguarding user privacy remains a non-negotiable priority.

In its statement, OpenAI signaled a major shift in its long-term privacy roadmap, previewing features designed to protect user conversations even from itself:

Together, these steps suggest OpenAI is reshaping its privacy framework to lock user conversations behind far stricter controls — reserving any human access for only the most exceptional, safety-critical situations.

The Times lawsuit is one of the most advanced cases in a series of copyright challenges facing OpenAI, Microsoft, Google, Anthropic, and others. A ruling requiring the release of millions of user logs could influence:

In many ways, this case is laying the groundwork for how copyright, privacy, and AI development will coexist in the years ahead.

Q: Why does The New York Times want access to 20 million ChatGPT logs?
A: The Times argues it needs a sample of conversations to determine whether ChatGPT reproduces or relies on its copyrighted reporting. The sample is meant to reveal how users interact with the model and whether any outputs closely match Times journalism.

Q: If the chats are anonymized, why are people still uncomfortable?
A: Even when names and personal details are removed, the content of a conversation can still feel sensitive or private. Many users do not expect their AI interactions — which often include medical questions, emotional issues, or personal tasks — to ever be reviewed by outside attorneys in a legal case they aren’t involved in.

Q: Why did the judge approve the request?
A: The judge determined that privacy risks were sufficiently addressed through strict protective orders, de-identification, and secure review environments. From the court’s perspective, these measures make the request appropriate for a copyright discovery process.

Q: What is OpenAI’s main concern?
A: OpenAI argues that the scale of the sample (20 million conversations) and the sensitivity of user content create risks to user trust — even with privacy protections in place. The company believes that more targeted alternatives would have been enough.

Q: Is either side “wrong” in this dispute?
A: Not necessarily. The Times is following established discovery practices for copyright cases, while OpenAI is defending user privacy expectations at a scale no previous AI company has had to navigate. Their goals reflect different professional traditions: journalism’s emphasis on transparency versus AI’s emphasis on confidentiality.

This dispute reveals a deeper tension that now sits at the center of AI’s evolution. In approving the request for 20 million anonymized chats, the judge concluded that the existing protections were sufficient: the conversations would be scrubbed of identifying details, reviewed only inside ultra-secure environments, and covered by a strict federal protective order. From the court’s perspective, these measures reduce privacy risks to an acceptable level for a copyright case.

At the same time, OpenAI’s objections reflect concerns that extend far beyond this lawsuit. The company argues that the scale of the request — tens of millions of conversations across two years — captures a vast amount of personal, sensitive material that users never expected to enter litigation.

Even without names attached, the content itself can reveal relationships, health issues, financial details, and other private aspects of someone’s life. That sensitivity, OpenAI says, makes the request unusually intrusive and potentially damaging to long-term user trust.

Both sides have legitimate arguments.

The New York Times is “right” that broad discovery requests are standard in copyright cases; anonymization reduces risk; protective orders restrict misuse; and plaintiffs must be able to examine evidence to prove whether an AI system replicates their journalism.

OpenAI is “right” that users expect a high degree of confidentiality; the volume of data involved is massive; the contents of chats can remain personal even after de-identification; and that maintaining user trust is crucial for AI products used by hundreds of millions of people.

The clash is not just legal — it’s philosophical.

Journalism has historically championed transparency, accountability, and public scrutiny.
AI systems are built on expectations of privacy, encryption, and user confidentiality.
These traditions collide here in a way that has no easy or comfortable answer.

For the public, this raises a straightforward but important question:
How private are AI conversations in practice, especially when legal disputes arise?

Many users are uncomfortable with the idea that anonymized versions of their chats — even stripped of names — could be reviewed by outside attorneys for a lawsuit they have no part in. Others accept it as a necessary step in determining whether AI systems rely too heavily on copyrighted material.

Ultimately, this case is about more than The New York Times or OpenAI. Its outcome will influence:

In many ways, this lawsuit is defining the early boundary lines of how copyright, privacy, and AI development will need to coexist in the years ahead.

Editor’s Note: This article was created by Alicia Shapiro, CMO of AiNews.com, with writing, image, and idea-generation support from ChatGPT, an AI assistant. However, the final perspective and editorial choices are solely Alicia Shapiro’s. Special thanks to ChatGPT for assistance with research and editorial support in crafting this article.