Generative AI tools like ChatGPT, Claude, Harvey, and Microsoft Copilot are increasingly being used to research legal questions, review documents, summarize information, and draft content. While these tools can be helpful, their use can also create unexpected risks to attorney‑client privilege and work product protections if not used carefully.

Recent court decisions have begun to address these risks, and they have not all reached the same result.

United States v. Heppner (S.D.N.Y.) — Privilege Denied

In U.S. v. Heppner, a criminal case, a federal court held that documents a defendant created using a consumer AI platform—on his own initiative, rather than at the direction of his attorney—were not protected by either the attorney client privilege or the work product doctrine. The court emphasized that the platform’s terms permitted the provider to retain user inputs and to disclose them to third parties, including the government, and it suggested the result might have been different had counsel directed the use through a tool with appropriate confidentiality protections.

Warner v. Gilbarco, Inc. (E.D. Mich.) — Work Product Protected

Days later, in a civil case, Warner v. Gilbarco, Inc., a different federal court reached the opposite result on work product, holding that a self represented litigant’s use of a public AI tool to prepare her own filings did not waive that protection because generative AI is a tool rather than a third party recipient and work product is waived only by disclosure to an adversary.

These cases arose in different procedural settings—one criminal, one civil—and turned heavily on their particular facts, including whether the AI use was directed by counsel. Together, they signal a quickly developing and unsettled area of law. Other courts, applying different state or circuit rules, may take a more aggressive view and treat the act of entering information into an AI tool as a waiver of privilege or work product protection. Because a waiver, once it occurs, is often irreversible and can reach beyond a single document, the prudent course is to use AI tools with caution because of these risks.

Understanding Why These Rules Create Immediate Risk for Businesses

Many publicly available AI tools operate in ways that can expose or retain information provided by users. When sensitive or confidential information is entered into these systems:

• The information may be stored, reused, or accessible beyond the user’s control
• It may be considered disclosed to a third party, potentially waiving privilege
• AI‑generated outputs may be treated as discoverable material – meaning that a party to litigation may be required to produce AI generated documents, including documents disclosing legal strategies

These risks are particularly relevant as clients increasingly turn to AI tools to independently:

• Research legal questions
• Analyze legal documents
• Validate or “double check” legal advice

How AI Tools Can Expose Privileged and Confidential Information

Most client use of AI has nothing to do with litigation or with seeking legal advice: drafting, summarizing, and analyzing information in the ordinary course of business. Those interactions are ordinary business records. Like emails, text messages, and chat logs, AI prompts and the responses they generate are generally discoverable, and they carry no privilege to begin with. Clients should not assume that AI interactions unrelated to litigation or in lieu of asking a licensed attorney for legal advice are shielded from discovery, or that the line between “business” use and “legal” use is a clean one. As Heppner illustrates, even AI use undertaken for a legal purpose may not be protected.

AI use can also create preservation obligations. Once litigation is reasonably anticipated, AI prompts and outputs may be subject to a litigation hold. Because that data often resides with a third party provider rather than on the client’s own systems, preserving and producing it can be more complicated than with ordinary documents—and the automatic deletion of chat histories can create risk of a spoliation claim against the user.

Key Considerations for Clients to Protect Privilege and Confidentiality

To help protect privilege and confidentiality:

• Avoid entering sensitive or confidential information into publicly available or consumer AI tools
• Treat AI platforms as third‑party environments unless appropriate safeguards are clearly in place
• Recognize that use of AI does not automatically preserve privilege or work product protections

How Ice Miller Approaches AI Responsibly

Ice Miller takes a controlled and governed approach to the use of artificial intelligence to protect our clients while delivering the efficiency benefits AI can offer.

• AI as an assistive tool, not a substitute — AI supports legal analysis but is not a substitute for the professional judgment of our attorneys
• Rigorous evaluation of all tools — Every AI tool used in connection with client matters undergoes thorough review for:
  • Security and data protection
  • Confidentiality safeguards
  • Accuracy and reliability
• Strict information controls — We implement strict controls designed to protect client information and preserve privilege and confidentiality at every stage
• Attorney accountability — Attorneys remain fully responsible for all work product and client advice
• Client consent — When appropriate, we obtain client consent before using AI tools in matter workflow

How to Use AI Safely in Connection with Legal Matters

If you are considering using AI tools in connection with legal matters, we recommend:

• Consulting with your Ice Miller attorney first before using AI on any sensitive or privileged matters
• Coordinating on appropriate tools and safeguards — Not all AI platforms are created equal, and the choice of platform and the subscription level can determine whether privilege is preserved
• Leveraging firm‑supported solutions where confidentiality and privilege protections are in place

Ice Miller can work with you to safely incorporate AI into workflows, using tools and environments designed to protect your information while still delivering efficiency and value.

Questions or Guidance on AI and Privilege

If you have questions about the use of generative AI and how it may affect privilege in connection with your matters or would like guidance on selecting and using generative AI tools safely and appropriately, please contact your Ice Miller attorney.

This publication is intended for general information purposes only and does not and is not intended to constitute legal advice. The reader should consult with legal counsel to determine how laws or decisions discussed herein apply to the reader's specific circumstances.