Federal Court Rules that a Defendant's Communications with Generative AI are Not Privileged

A federal judge in the Southern District of New York ruled that written exchanges between a defendant and a generative artificial intelligence platform are not protected The decision addressed the case of United States of America v. Heppner, in which an individual facing fraud charges used the platform Claude to process information learned from counsel and to generate documents for discussions with counsel. The defendant shared the resulting documents with counsel, but the interactions occurred without direction from the attorney. FBI agents seized 31 documents memorializing the exchanges during a search of the defendant's home. The court rejected privilege claims for several reasons. The documents were not communications between the defendant and counsel. The exchanges also lacked confidentiality because users of publicly accessible AI platforms do not hold substantial privacy interests in such conversations, and the platform's policy permits disclosure to third parties. The court further found that the defendant did not engage with the AI for the purpose of obtaining legal advice. The ruling similarly denied work product protection. The documents were not prepared