Court Finds AI Platform Communications Not Protected by Privilege | U.S. v. Heppner
Summary
U.S. v. Heppner, No. 25-cr-00503 (S.D.N.Y. 2026): Previously discussed in Akin alert: SDNY Rules Communications with a Public Generative AI Platform Are Not Protected by Attorney-Client Privilege or Work Product Doctrine
As part of defendant’s arrest for securities and wire fraud, law enforcement officials executed a search warrant at defendant’s home, seizing numerous documents and electronic devices. Among the seized materials were communications the defendant had with the generative AI platform, Claude, in which Heppner asked it to prepare reports outlining defense strategies. Heppner used Claude to prepare this information after he received a grand jury subpoena and, importantly, did so without the direction of his counsel.
Heppner asserted both attorney-client privilege and the work product doctrine, but the court found that neither protected the materials from discovery. With respect to attorney-client privilege, the court first noted that since Claude is not an attorney, these materials do not constitute communications between the defendant and his counsel. Next, the materials were not deemed confidential as Claude is a third-party platform outside the attorney-client relationship and Claude’s privacy policy explicitly provides that users’ inputs/outputs are used to train the model as well as are disclosed to third parties. Finally, Heppner’s use of Claude was not for the purpose of seeking legal advice because his counsel never directed him to create the materials at issue. The court similarly denied Heppner’s work product argument because, even assuming they were created in anticipation of litigation, the materials were not prepared by or at the direction of the defendant's counsel. Nor did the materials reflect his counsel’s strategy at the time of their creation. Currently, the case is at the trial stage. (See Memorandum, No. 25-cr-00503 (S.D.N.Y.), Dkt. 27.)
