Today’s artificial intelligence (AI) platforms have shown impressive capabilities that keep evolving. As those capabilities improve, tax departments may be inclined to leverage AI’s analytical power. While this technology has led to incredible efficiencies, we’ve been warning in-house tax departments about which platforms they use and what information they share to avoid waiving privilege or creating documents that cannot be protected.

The US District Court for the Southern District of New York recently held that a criminal defendant’s communications with a publicly open AI platform were not protected by attorney-client privilege nor the work product doctrine. United States v. Heppner, No. 1:25-cr-00503-JSR (S.D.N.Y. Feb. 17, 2026).

In Heppner, the defendant transmitted confidential information to a public AI system and generated various documents that incorporated the information. The district court concluded that the documents created were not protected by attorney-client privilege nor the work product doctrine. The AI platform was a public and open system that did not provide confidentiality, the documents created by the AI platform were not prepared by or at the direction of counsel (in fact, counsel had no idea the client was using AI), and documents (regardless if created by AI or other means) can never be cloaked with privilege simply because they are later sent to a lawyer.

Practice point:

While the facts of Heppner are probably distinguishable from how tax departments typically use AI platforms, the case serves as a reminder for tax professionals to have good hygiene when using them. All in-house tax professionals should exercise caution when inputting confidential information into AI platforms and, where possible, rely on closed, internal AI systems that are only accessible by relevant persons within their corporation. Even then, Heppner makes clear that AI platforms are not lawyers, and disclosures of privileged information to such platforms risk waiving that privilege.

Given this risk, tax professionals should use great caution when using AI for sensitive legal issues. At a minimum:

• Do not input any sensitive, confidential, or privileged information into publicly open AI systems.

• Remember that AI is not a lawyer, so asking AI legal questions is not the same as asking a lawyer for legal advice.

• For tax issues that are likely to result in a contentious audit or litigation, work with in-house or outside counsel to establish best practices on AI use to maximize attorney-client privilege and work product protection.