AI Privilege Issues Get More Complex as New Courts Weigh in

Courts across the U.S. are weighing in on evidentiary issues related to AI. Key among these is whether prompts entered into commercial AI models and the subsequent outputs are protected by evidentiary privilege. We previously wrote about a groundbreaking case out of New York, United States v. Heppner. In that case, the court ruled that AI prompts and outputs were not protected by attorney-client or work product privilege. This is because the defendant in Heppner used a commercial AI model and sent the outputs to their lawyer. Since commercial AI developers had access to the prompts and outputs, the court ruled that there was no reasonable expectation of privacy. A recent Norton Rose Fullbright memo examines Heppner in the context of subsequent cases that diverge on this point:

“Both Warner and Morgan decisions also depart from Heppner by noting that GenAI is a ‘tool,’ not a person that would raise potential third party waiver considerations. Relatedly, the Morgan court appears to take a different position than the Heppner court regarding whether the use of a publicly available GenAI tool by itself eliminates any expectation of privacy the user has in the tool such that privilege protections are automatically waived. The Morgan court raises an analogue in an individual’s use of email, citing prior circuit decisions holding that the mere fact that information is held by a third-party intermediary, like an internet service provider, does not extinguish a user’s expectation of privacy and confidentiality.”

So far, at least two courts have differed from the outcome in Heppner. The judge in one of these cases compares commercial AI use to email. Noting that even though a service provider is technically in possession of such communications, that does not void the reasonable expectation of privacy. So far, we only have trial court rulings to go on. As this issue increasingly crops up in cases around the country, we can expect appellate courts to start weighing in.