Generative AI and Privilege: Practical Lessons from Two Early Decisions and What Comes Next

Two federal district courts addressed generative AI in the privilege context in February 2026. The decisions applied existing attorney-client privilege and work product doctrines to new factual settings without announcing any shift in the law. In United States v. Heppner, the United States District Court for the Southern District of New York denied privilege and work product protection. A financial services executive had generated legal strategy materials using a generative AI tool without counsel direction. The court found the materials were not confidential under the platform terms, were not communications with an attorney for legal advice, and were not prepared at the direction of counsel or reflective of attorney mental impressions. In Warner v. Gilbarco, the United States District Court for the Eastern District of Michigan upheld work product protection. The case involved a pro se litigant's AI-assisted analysis prepared after the close of discovery. The court determined the materials reflected the plaintiff's own mental impressions in anticipation of litigation, with the plaintiff effectively serving as his own attorney. Disclosure to the public AI platform did not waive protection because it did not meaningfully increase the likelihood that an adversary would receive the material, and because generative AI programs are tools, not persons. The rulings reflect continuity in privilege doctrine. Organizations should inventory generative AI tools in use, evaluate their confidentiality and data-handling terms, and prohibit use for legal analysis or strategy without counsel involvement and direction. Existing litigation hold and preservation protocols should also account for AI-generated materials, including prompts, outputs, and related metadata.