Federal Court Rules that a Defendant's Communications With Generative AI are Not Privileged
In a question of first impression, a federal judge in the Southern District of New York agreed with the United States government that certain written exchanges between a defendant and a generative artificial intelligence platform were not protected from government inspection by either the attorney-client privilege or the work product doctrine.
A recent federal court decision addressed the intersection between generative artificial intelligence ("AI") and privilege claims.
In United States of America v. Heppner, an individual facing fraud charges had: (i) input into the generative AI platform "Claude" information that he learned from counsel; (ii) used AI to create documents for the purpose of speaking with counsel to obtain legal advice; and (iii) subsequently shared the contents of the AI documents with counsel. (The defendant did this on his own to facilitate the provision of legal advice, but not at counsel's direction.)
FBI agents executed a search warrant at the defendant's home and seized material, including 31 documents memorializing the communications with Claude. According to defense counsel, the documents contained reports that outlined defense strategy and what the individual might argue in response to the government's charges.
Heppner attempted to bar the government from using the communications, arguing that the attorney-client privilege and work product doctrine should apply. The government disagreed. On February 17, 2026, the court issued its written decision agreeing with the government's position.
The court rejected privilege protection for multiple reasons. First, the AI documents were not communications between Heppner and his counsel. Second, the court found that the communications were not confidential because "AI users do not have substantial privacy interests in their conversations with" a publicly accessible AI platform, and Claude's privacy policy reserves the right to disclose data to third parties. Third, the court found that the individual did not communicate with Claude for the purpose of obtaining legal advice because he was not directed to do so by counsel.
Similarly, the court rejected application of the work product doctrine where the AI documents were not prepared by or at the behest of counsel and they did not reflect defense counsel's strategy at the time they were created.
In light of Heppner, litigants should exercise caution when using AI in legal contexts, and ensure consultation with counsel when considering whether to do so.
