More of our clients are turning to ChatGPT, Claude, Copilot, and similar tools to research legal questions, summarize documents, and “double check” their lawyer’s advice. While this is understandable, you may be surprised to learn that what you type into an AI tool, and what it types back, likely isn’t protected by the attorney-client or work product privileges. And uploading your lawyer’s emails, memos or call notes into one of these tools can put the applicable protection of those materials at risk. The law is new and still developing, but the early signals are clear enough that every client should know the risks.
Attorney-Client and Work Product Privileges
The Attorney-Client Privilege
Indiana protects confidential communications between you and your lawyer by statute. Under Ind. Code § 34-46-3-1, attorneys cannot be forced to testify about confidential communications made to them in their professional work, or about the advice they gave. To claim the privilege, you must show two things: that an attorney-client relationship existed, and that the communication was confidential and made to get legal advice. The key word is confidential. If someone outside the attorney-client relationship is in the room, on the call, or copied on the email, the privilege usually doesn’t attach.
Work Product Protection
The work product privilege is a separate protection. It covers materials prepared because a lawsuit is coming or already underway, such as a lawyer’s notes, case analysis, and strategy memos. Under Indiana Trial Rule 26(B)(3), the other side can get these materials only by showing a substantial need and that it can’t get the same information another way. Even then, a lawyer’s mental impressions, opinions and legal theories stay off-limits.
Your Lawyer Must Guard Your Secrets, But You Hold the Key
Indiana lawyers are required by law to protect client confidences. Ind. Code § 33-43-1-3 says an attorney shall “[m]aintain inviolate the confidence and, at every peril to the attorney, to preserve the secrets of the attorney’s client.” But the privilege belongs to you, the client, and it can be waived by your own conduct. You can give it up on purpose, or by accident. Forwarding your lawyer’s email to a business partner, uploading your lawyer’s advice into an AI tool to help prepare a board deck or talking about your case in front of people who don’t need to be there can all do it. Once waived, the protection is usually gone for good, and it can spread to related communications.
Where AI Use Puts Privilege at Risk
What We’re Seeing
In the world of implementing AI tools into business practices, client disclosure of legal information usually falls into two scenarios. The first scenario is clients asking AI tool legal questions, either their own or right after talking with their lawyer, to better understand what was discussed or to “double check” the lawyer’s advice.
The second scenario is clients uploading materials from their lawyer, such as memos, email exchanges or transcripts of calls, and asking the AI tool to summarize, explain something, draft a response, or create a presentation with this information. Both patterns can lead to the same problem: sharing privileged information with an outsider.
Asking AI Legal Questions
The general rule is simple. You waive the privilege if you voluntarily disclose, or consent to disclosure of, any significant part of a privileged communication.
In February 2026, a federal court in New York became one of the first courts to apply these principles to AI usage in United States v. Heppner1. After learning he was the target of a criminal investigation, Mr. Heppner used Claude on his own, without his lawyers’ direction, to work through his defense strategy. He fed in information he had learned from his lawyers and later shared the AI’s output with them. The court ruled that none of it was privileged for three reasons:
- Claude is not a lawyer. There was no attorney-client relationship with the AI.
- The conversations weren’t confidential. The platform’s privacy policy allowed the company to collect what users typed, use it to train its models and share it with others, including the government.
- He wasn’t seeking legal advice from the AI. It didn’t matter that he later passed the results along to his lawyers.
This decision is a direct warning to anyone who chats with AI after meeting with their lawyer. Typing your lawyer’s advice into a chatbot to “check” it doesn’t create a privileged communication. But, it may mean you’ve told an outsider what your lawyer told you.
Uploading Your Lawyer’s Emails, Memos, and Call Notes
What happens when you upload an existing privileged communication or documentation into an AI tool?
Information prepared in the normal course of business and not “in anticipation of litigation” does not enjoy work product protection. Most transactional and deal advice falls into this category, so any protection from disclosure rests solely on the attorney-client privilege. Confidential communications between you and your lawyer, such as internal memoranda or draft documents, will likely still be protected by that privilege, but once they are shared outside the attorney-client relationship, including with an AI tool, that protection may be waived and the materials could be discoverable if compelled.
However, if you are under the threat of litigation or in active litigation, sharing it with a third-party does not automatically waive that protection like it would with the attorney-client privilege. Generally, waiver occurs only when the material is disclosed to an opponent, or in a way that makes it likely to reach one.
In February 2026, a federal court in Michigan2 held that information disclosed by a pro se party to ChatGPT was her protected work product and that her use of ChatGPT during a case did not waive that protection. The court cited to precedent that said the work product waiver has to be to an adversary or in a way likely to get in an adversary’s hands. AI programs “are tools, not persons even if they may have administrators somewhere in the background.” Nevertheless, courts have found this ruling does not provide for a blanket objection to discovery3. Courts must still “examine the purpose and circumstances of the particular communications rather than presume protection[.]” Id. Independently generated communications are not transformed into attorney work product merely because the client is a litigant or because the resulting content may bear on the action.
Indiana Courts Apply A Balancing Test
In Indiana, courts decide whether a disclosure waived the privilege through a balancing test by weighing “the reasonableness of the precautions to prevent inadvertent disclosure, the time taken to rectify the error, the scope of discovery,” the extent of the disclosure, and “an overreaching issue of fairness and the protection of an appropriate privilege which, of course, must be judged against the care or negligence with which the privilege is guarded with care and diligence or negligence and indifference4.” This approach allows for fair consideration of the circumstances attendant to the disclosure without placing an onerous burden on the holder of the privilege or encouraging carelessness.
In JWP Zack, privileged documents were accidentally mixed into a large document production. Because the lawyer caught it right away, the privilege survived.
In Buntin5, a doctor put privileged material, including communications with his lawyers and his notes from a meeting with his lawyer, into a file at a shared office. His officemate then reviewed that file to prepare for a deposition. The court found the privilege waived because the doctor deliberately placed the documents where someone else would use them and failed to protect them.
In comparing these examples, uploading a lawyer’s memorandum into a consumer AI tool feels more like Buntin than JWP Zack because it is deliberate and puts the material in the hands of a company outside your control. Using a lawyer-approved, enterprise-grade tool with strong confidentiality terms may provide a better argument for maintaining privilege, but it’s not a guarantee.
Practical Tips for Clients Who Use AI
Client use of AI isn’t going away. Our goal is to ensure clients use these tools wisely and think about the consequences of what information is being provided and how it is stored. A few tips:
- Ask your lawyer, not the chatbot, about your case.
- Don’t paste your lawyer’s emails, memos or call notes into a public AI tool, especially on sensitive deal topics like earnouts or potential indemnification claims, which are among the most commonly litigated issues in transactions.
- Don’t use AI to “double check” your lawyer’s advice. If you have questions, call us.
- Don’t let AI note-takers record or transcribe calls with your lawyer unless your lawyer approves the tool and how the information is maintained.
- Treat every AI platform as an outsider unless your lawyer has confirmed it has proper safeguards.
- Use enterprise-grade or lawyer-approved tools and turn off model training and chat history where you can.
- If AI would help with your legal matter, ask your lawyer first.
- Keep everyday business use of AI separate from legal matters. Ordinary AI chats are usually discoverable.
- Don’t delete AI chats once a dispute is brewing, such as a disagreement over an earnout payment or an indemnification claim after closing. You may be required to preserve them.
- Never upload documents covered by a protective order or confidentiality agreement. This includes a counterparty’s data room materials, since uploading them could breach your NDA or letter of intent.
- If you think you’ve shared something privileged, tell your lawyer right away. Acting quickly matters.
- Businesses: adopt a written AI policy and train your team on it.
AI tools can save time, but the law hasn’t caught up to them yet. For now, the safest assumption is that a public chatbot is an outsider, and sharing your lawyer’s advice with it may mean losing protection you can’t get back. Your lawyer is bound to guard your secrets “at every peril,” but only you can avoid giving them away. When in doubt, talk to your lawyer before you talk to the chatbot.
This post is for general information only and is not legal advice. Contact Gutwein Law with questions about your situation.