Resisting the New York Times' encroachment on user privacy

Meta description: OpenAI is pushing back against The New York Times' request for 20 million private ChatGPT logs while adding new privacy and security tools for users.

Tags: OpenAI, ChatGPT, privacy, New York Times lawsuit, data security

Featured image: A friendly abstract illustration of a shield made of soft geometric shapes surrounding a chat bubble, in warm blues and oranges.

If you use ChatGPT for personal projects, work drafts, or everyday questions, two recent developments from OpenAI matter to you. The company is fighting a court order that would force it to hand over 20 million private ChatGPT conversations to The New York Times, and at the same time it is speeding out new privacy and security features designed to keep your chats safer. Here is what is happening, why it matters, and what practical steps you can take today.

The lawsuit battle over your chat logs

The dispute goes back to the copyright case The New York Times filed against OpenAI in December 2023, accusing the company of using Times articles without permission to train its AI models. In May 2025, the court ordered OpenAI to preserve and produce ChatGPT conversation logs, including chats users believed were deleted, as potential evidence.

OpenAI says that request is far too broad. In its appeal filed in November 2025, the company argued that demanding roughly 20 million conversations sweeps in an enormous amount of private material — from sensitive personal questions to confidential business information — that has nothing to do with the copyright claims. OpenAI has asked the court to narrow the scope significantly, suggesting smaller, anonymized sample sets instead.

The outcome is still pending, so it remains uncertain how much data, if any, will ultimately be produced. OpenAI has committed to opposing any broad production of user chats and has criticized the demand as a privacy overreach that breaks the trust users place in the service.

Why this matters even if you have nothing to hide

Most people treat ChatGPT a bit like a private diary or a trusted colleague. They type out legal questions, health concerns, relationship issues, or half-finished business plans. If courts can compel a company to hand over millions of those conversations, the expectation of privacy changes for everyone — not just the small subset that might ever be relevant to a lawsuit.

There is also a chilling effect to consider. If people start worrying their chats could end up in a courtroom file, they may hold back from using AI tools for genuinely useful things, or they may sanitize their questions in ways that make the tools less helpful.

To be clear on the facts: this is a dispute over evidence in a copyright case, not an accusation that OpenAI is selling your chats. But the legal fight has pushed privacy and security to the top of OpenAI's priority list, and the company has responded with a wave of new protections.

The new privacy and security features OpenAI rolled out

OpenAI has announced several changes, some directly in response to the legal battle:

• Encrypted, retroactive retention options. OpenAI has committed to challenging the data preservation order while also exploring stronger encryption for conversations and more control over how long chats are stored. Enterprise and education customers can already control retention periods.
• Temporary chats. ChatGPT now offers a temporary chat mode that does not save conversations to your history, does not use them for training, and does not retain them long term — useful for sensitive one-off questions.
• Easier opt-out of training. Free, Plus, and Pro users can turn off "improve the model for everyone" in their data controls, keeping their chats out of future model training.
• Improved age prediction. ChatGPT is rolling out an age-prediction system to route suspected underage users to a more restricted experience, part of a broader safety push.
• Sensitive-content guardrails. OpenAI is adding stronger safeguards around conversations involving self-harm and other sensitive topics, including better crisis resources.

The company has framed these changes as part of its principle that personal conversations with AI should be treated as private by default.

What you can do right now to protect your chats

You do not have to wait for the lawsuit to resolve. A few quick settings changes go a long way:

1. Turn off training. Go to Settings > Data Controls and switch off "Improve the model for everyone" if you have not already.
2. Use temporary chats for sensitive topics. Click the chat model picker and select temporary chat when discussing anything personal or confidential.
3. Review and delete old conversations. Browse your history and remove anything you would not want surfaced later.
4. Check your enterprise settings at work. If your company has a ChatGPT Team or Enterprise plan, ask your admin what retention rules are in place.
5. Avoid sharing what you would not share elsewhere. Treat any cloud service, including ChatGPT, with the same caution you would apply to email.

What happens next

The appeal process will take time, and the courts have not yet ruled on OpenAI's request to narrow the order. Whichever way it goes, the case is shaping up to be a landmark moment for AI privacy — a test of how the legal system balances evidence collection against the privacy expectations of millions of everyday users. Meanwhile, OpenAI's rapid rollout of new controls gives users more say over their data than they had a year ago.

The practical takeaway: keep an eye on the case, but more importantly, take ten minutes today to review your ChatGPT privacy settings. The tools are there — you just have to use them.