AI and Legal Ethics: Privilege, Discovery, and Risk

This Law Practice Today podcast episode discusses legal ethics and privilege issues arising from clients’ and lawyers’ use of generative AI, featuring consultant and ethics attorney Jennifer Ellis. Ellis compares two February 10, 2026, bench rulings—United States v. Heppner (S.D.N.Y.) and Warner v. Barco (Michigan)—that reached opposite conclusions on whether a client’s AI interactions are discoverable, highlighting differences in judges’ approaches and understanding of AI privacy policies and training. She advises firms to warn clients via engagement agreement clauses about unclear privilege, discoverability risks, and billing impacts when clients send AI-generated materials. The conversation also covers widespread sanctions for AI-fabricated citations and misstated holdings, urging lawyers to take ethics training, promptly correct mistakes, and verify citations by reading cases using tools like law libraries or Google Scholar.00:00 AI Fabrication Trouble00:26 Show Intro Disclaimer01:05 AI Privilege Cases02:04 Guest Background04:37 Heppner vs Warner07:21 Privacy Policies Training10:44 Client Warnings Clauses15:34 Billing Value Added20:49 Hallucinations Sanctions26:02 Mitigating AI Risks27:02 Warner Ruling Explained28:08 Policies Resources Wrap29:16 Final Thanks Outro