Why Judges Are Banning AI Tools in Discovery: The Privacy Risks You Need to Know

In recent months, a growing number of judges have started explicitly prohibiting the use of public AI tools—such as ChatGPT and similar services—during the discovery phase of litigation. The rulings are being driven by a clear concern: the risk that sensitive, confidential information will be leaked outside the legal process. For anyone involved in a lawsuit, or simply worried about the fate of their personal data, these decisions deserve attention.

What Happened

In June 2026, Bloomberg Law reported on a series of rulings where judges barred parties from using public generative AI services for discovery work. The orders typically cite the risk that data submitted to these platforms could be used for model training, inadvertently disclosed, or intercepted. While the rulings vary, a common thread is the recognition that third-party AI services do not offer the confidentiality guarantees that legal proceedings require.

Separately, in a January 2026 article, Bloomberg Law detailed how AI notetaking tools can create wiretapping and discovery pitfalls. Those concerns extend to any AI service that records, transcribes, or processes conversations that may contain privileged or proprietary information.

No single jurisdiction has issued a blanket ban; rather, individual judges are acting on a case-by-case basis when they see a concrete risk. The trend is still emerging, and it remains unclear whether appellate courts or rule-making bodies will formalize guidance.

Why It Matters

For most people, the legal system already feels opaque. The arrival of AI in discovery adds another layer of risk that isn’t always obvious. Here’s what’s at stake:

  • Data exposure. When a lawyer or a party uploads a deposition transcript, a contract, or a list of trade secrets to a public AI tool, that data may be stored on servers outside the control of the court. Even if the tool’s privacy policy claims not to use the data for training, few services offer binding contractual guarantees of deletion or isolation.
  • Loss of privilege. Attorney-client privilege and work‑product protections depend on maintaining confidentiality. Sending a document through a public AI service could be seen as waiving that privilege—potentially devastating in litigation.
  • Wiretapping risks. AI notetaking tools that listen in on meetings or calls, even with consent, can capture discussion that should never leave the room. If those recordings are later discoverable, the consequences can be severe.
  • Erosion of trust. When parties suspect that the other side is using AI tools that may not be secure, cooperation in discovery suffers. Motions to compel and protective orders multiply, driving up costs and delay.

These risks aren’t hypothetical. The rulings show that judges are already taking action. For consumers, the lesson is that your personal information—medical records, financial documents, private messages—could end up in front of an AI model if your case involves any digital discovery.

What You Can Do

If you are party to a lawsuit or expect to be involved in one, there are practical steps to protect your data:

  1. Ask about AI use early. In your first meeting with counsel or during the initial case conference, ask explicitly whether any party intends to use public AI tools for discovery. You have a right to know what safeguards are in place.
  2. Require a protective order. Request that the court enter an order barring the use of public AI services for any discovery material that contains personal or confidential information. Some judges are already including such language; you can propose it.
  3. Use court‑approved or enterprise‑grade tools. Many law firms have access to secure AI platforms that are contractually bound to keep data confidential and are not trained on user inputs. Insist that these be used instead of free public tools.
  4. Review privacy policies—and don’t stop there. Even if a tool claims not to store your data, look for independent audits or certifications (e.g., SOC 2, ISO 27001). A well‑written policy is no substitute for enforceable commitments.
  5. Educate yourself on wiretapping laws. If AI notetaking is used in any meeting or telephone call, ensure that all participants consent and that the recording is stored in a way that preserves privilege. Better yet, use a dedicated, encrypted transcription service that does not rely on cloud‑based AI.

Broader Implications

These judicial bans are part of a larger conversation about how AI fits into the legal system and into everyday life. The same risks that trouble judges in discovery—data leakage, lack of transparency, loss of control—apply to how consumers use AI for personal tasks. If courts are uncomfortable trusting these tools for legally protected information, individuals should be equally cautious about using them for tax returns, health records, or private correspondence.

The trend is likely to accelerate. As more rulings emerge, we may see state and federal rules committees issue formal guidance. In the meantime, the message from the bench is clear: privacy comes first, and AI tools that cannot guarantee it have no place in the courtroom.

Sources

  • Bloomberg Law, “Judges’ Public AI Bans During Discovery Zero in on Privacy Risk,” June 5, 2026.
  • Bloomberg Law, “AI Notetaking Poses Wiretapping, Discovery, and Ethical Pitfalls,” January 28, 2026.