What Privilege? How AI Interactions Are Discoverable.
Earlier this year in a matter of first impression, the United States District Court for the Southern District of New York in United States v. Heppner, 820 F.Supp.3d 292 (2026) held that an AI user’s communications in connection with a pending criminal investigation were not protected by the attorney-client privilege or the work product doctrine. In that case, Heppner, a target defendant, used Claude AI to prepare a defense strategy. Later, the government sought access to Heppner’s interactions with Claude. Heppner, through his counsel, asserted privilege over the documents claiming, among other things, that Heppner had input things into Claude that he learned from his counsel for the purpose of speaking with counsel to obtain legal advice.
The Court quickly determined that the documents were not communications between Heppner and his counsel. The Court noted that, of course, Claude is not an attorney and is really akin to the use of other internet based software. The Court went on to say that AI users do not have substantial privacy interests in conversations with publicly available AI platforms.
We know there is a temptation even for clients of law firms when a legal problem, contract or transaction arises to “run it through AI.” The problem is when doing so, the resulting interaction is very likely not protected by the attorney-client privilege.