Lions and Tigers and AI, Oh My! A Cautionary Tale on Privilege
Long, long ago, in the early 2000s, legal research meant flipping through casebooks, statutes, and printed binders. Today, clients and attorneys can access legal information in seconds thanks to Artificial Intelligence. AI tools can quickly summarize facts, test arguments, and provide background information before an attorney is ever consulted. While AI tools can be useful, a federal court recently ruled that confidentiality is not guaranteed when using such platforms.
On February 10, 2026, a federal judge in the Southern District of New York ruled that documents generated by a litigant using a commercial AI platform were not protected by the attorney-client privilege or the attorney work-product doctrine. Transcript of Pretrial Conference at 6, United States v. Heppner, No. 1:25-cr-00503-JSR (S.D.N.Y. Feb. 10, 2026). Judge Jed S. Rakoff ruled that AI-generated documents created by a defendant were discoverable. Id. The defendant had used an AI system to analyze legal issues related to his criminal case, which he then shared with his attorney. When the documents surfaced in discovery, the government argued they were not privileged, and the court agreed.
Attorney-client privilege protects communications between an attorney and a client made for the purpose of seeking or providing legal advice. Tex. R. Evid. 503. Because confidentiality is central to the ruling, the court focused on the fact that the AI provider’s terms disclaimed confidentiality. Tr. at 3. If a client prepares notes, summaries, or analyses independently, those materials generally remain discoverable and sending them to counsel after the fact is not a magical fix. In this case, the defendant created the documents on his own, without instruction from his attorney. Because the materials were not prepared as part of counsel’s litigation strategy, they were not protected.
This ruling has several impacts on municipalities, public officials, and private citizens. As AI tools become more accessible, it is tempting to use them before consulting legal counsel. However, entering sensitive facts or legal theories into commercial AI platforms may expose that information to discovery later. For Texas cities and other public entities, this issue is significant in matters such as personnel investigations, contract disputes, enforcement actions, and pending litigation. Staff members may believe they are being efficient by using AI to summarize documents or test legal arguments, but if those materials are created outside counsel’s direction and through third-party platforms, privilege may not apply. When legal issues are involved, AI tools should be used only with counsel’s knowledge and confidentiality safeguards.
Additionally, municipal officials and employees should be aware that even non-legal related uses of AI platforms may have unexpected consequences. Many governmental employees do not realize that uploading internal documents, correspondence, reports, or drafts for summarization, organization, or asking questions specific to a City project, employee, property, or enforcement matter, may later be subject to disclosure. Because AI platforms operate as third-party service providers that often disclaim confidentiality, there is no assurance that such communications will remain private. In reality, the use of non-confidential AI tools can create an unofficial record of the City’s internal deliberations, assumptions, and considerations that opposing parties may later seek to use in proceedings.
AI technology may be new, but the rules governing confidentiality and privilege are not. For public officials, city staff, and private clients, the lesson is a simple one: when in doubt, talk to your lawyer before sharing sensitive information with a stranger (AI) over the internet!
