Judges Are Banning Public AI in Court Cases — What That Means for Your Privacy at Work

Judges Are Banning Public AI in Court Cases — What That Means for Your Privacy at Work

A growing number of judges are issuing orders that prohibit lawyers and parties from using public artificial intelligence tools during discovery. The reason? Privacy risks that many professionals have not fully considered. If you use AI for document review, note-taking, or research with sensitive information, these rulings carry lessons that extend well beyond the courtroom.

What happened

In June 2026, Bloomberg Law News reported that several judges had begun to explicitly ban the use of public AI platforms — including large language models like ChatGPT, Claude, and Gemini — during the discovery phase of litigation. The orders cite concerns about data leakage, lack of confidentiality, and the fact that user inputs are often processed on third-party servers without guaranteed deletion.

A related article from January 2026 flagged a separate but overlapping issue: AI notetaking services can create wiretapping and discovery pitfalls, especially when used in meetings or calls that involve sensitive material. Taken together, these rulings signal a tightening of expectations around how professionals handle data when AI is involved.

The bans are not universal yet. They have appeared in federal and state cases, often in response to specific motions or as part of standing orders. But the pattern is clear: judges are paying attention to how AI tools process information, and they are not comfortable with the default privacy configurations of most public services.

Why it matters

The immediate impact is on legal professionals. If you are involved in litigation and use a public AI tool to summarize depositions, redact documents, or draft discovery requests, you could violate a court order — and potentially expose client data to third parties.

But the implications go further. Journalists handling confidential sources, researchers managing sensitive data, and knowledge workers in any industry that deals with proprietary or regulated information face similar risks. The same mechanisms that allow AI to generate useful answers — processing your inputs on remote servers, storing conversation history, and using data to train models — are what create exposure.

When you paste a draft affidavit into a public chatbot, you are effectively handing that text to a company whose privacy policy may permit use for model training or internal analysis. Even if the policy says the data is not retained, you rely on their word and their security. For court cases, that trust is no longer sufficient.

What readers can do

You do not need to stop using AI. But you need to make deliberate choices about which tools you use with sensitive information.

Start by reading the privacy policy and terms of service for any AI tool before you input anything confidential. Look for language about data retention, whether inputs are used for training, and whether the company offers enterprise versions with stronger protections.

For work that involves privileged, confidential, or personally identifiable information, use tools that promise zero data retention and are designed for professional use. Some providers offer dedicated enterprise tiers with contractual guarantees about data handling. Others operate locally on your machine, meaning nothing leaves your device.

If you are a legal professional, check your jurisdiction’s ethics rules and any standing orders in your active cases. Some bar associations have issued advisory opinions on generative AI. Even if no explicit ban exists, you can still be held responsible for inadvertent disclosure.

Concretely:

• Avoid pasting full documents, names, or case details into free, public AI chatbots.
• If you must use AI for analysis, use redacted versions or pseudonyms where possible.
• Keep a record of which AI tools you used and for what purpose, in case you need to explain your process later.
• For meetings and calls, disable AI notetaking features unless the tool explicitly stores data locally and you have consent from all participants.

Sources

• Bloomberg Law News, “Judges’ Public AI Bans During Discovery Zero in on Privacy Risk,” June 2026.
• Bloomberg Law News, “AI Notetaking Poses Wiretapping, Discovery, and Ethical Pitfalls,” January 2026.

These rulings are still new, and it is likely that more courts will follow. The broader lesson is that the privacy implications of using public AI tools are real, and they are being taken seriously by those who oversee the integrity of legal proceedings. Whether you work in law, media, or research, the same principles apply: know what your tools do with your data, and do not assume that a convenient tool is a safe one.