If you think this sounds self-serving, believe us, it is not. In a recent federal court case (United States v. Heppner), the judge required a defendant to turn over documents he generated using Anthropic’s Claude chatbot, rejecting the argument that the AI exchanges were protected by attorney-client privilege. In a separate ruling in a federal lawsuit in Michigan, the court treated certain AI chats as the litigant’s own work product. These rulings tee up the risk many take by sharing sensitive materials in an effort to engage AI platforms for legal research or analysis.
Business Takeaway: Employers should treat AI tools as useful productivity tools, not confidential advisors. If employees, executives, or in-house teams are using generative AI to analyze disputes, summarize legal advice, prepare investigation materials, or think through sensitive personnel issues, there is a risk that those inputs could later become discoverable or otherwise complicate privilege arguments. Now is a good time to revisit internal guardrails around AI use and to make sure legally sensitive facts and attorney communications are staying where they belong. Have you shared confidential information with AI chatbots, notetakers, or other applications? Is the AI notetaker on by default for every video conference?