AI tools are ‘like any other third party’, ‘not a licensed attorney’: Judge rules in Heppner case

AI/Artificial intelligence

• US court rules AI communications lack attorney-client privilege
• AI tools deemed third parties, not confidential in legal use
• AI data sharing may waive privilege, trade secret protection

Did our AI summary help?

A US federal court has ruled that communications with artificial intelligence tools are not protected under attorney-client privilege, treating such platforms as third parties. The decision came in United States v. Heppner, where Judge Jed Rakoff of the Southern District of New York examined whether interactions with a generative AI system could remain confidential.

Reuters has reported that in the case, the defendant argued that documents generated through interactions with Anthropic’s AI tool Claude were part of legal preparation and should be protected under attorney-client privilege and work product doctrine. The materials were said to include defence strategy prepared using information shared with legal counsel.

Judge Rakoff rejected the claim, stating that Claude is “not a licensed attorney” and cannot establish an attorney-client relationship. The court further held that the communications were “not confidential.” It added that even if some of the information shared with the AI tool was originally privileged, that protection could be waived once the data was entered into the system.

AI treated as third party

A central finding of the ruling is the classification of AI tools as external entities. The court said that sharing information with an AI platform is effectively the same as sharing it with “any other third party.” It noted that users consent to platform policies that allow data sharing with external entities, including regulators, weakening any expectation of confidentiality.

The ruling has implications beyond legal privilege. Confidentiality is also a requirement for trade secret protection, and courts may interpret sharing proprietary data with AI tools as a failure to maintain secrecy. As AI tools are increasingly used for drafting and analysis, the decision signals that existing legal standards will apply, treating such interactions as voluntary disclosure rather than protected communication.