Partner James T. Christie and Associate Danielle Izzo Mazzeo authored an informative article for the New York Law Journal entitled, “Drawing the Line Between AI and the Law.”  In the article, James and Danielle examine the growing use of AI as a legal resource and discuss how courts are increasingly finding AI chatbot prompts and responses discoverable.  They caution users to “think twice before asking your AI chatbot for legal advice as it could very well end up in the hands of your adversary.”

The authors examine recent cases addressing claims of privilege over AI-generated responses and documents.  In United States v. Heppner, the court held that neither attorney-client privilege nor work-product protection applied because the materials were not maintained as confidential or prepared by or at the direction of counsel.  The court's decision turned on three key issues: the defendant waived confidentiality by entering sensitive information into a third-party AI platform, the AI chatbot was not considered necessary to assist an attorney in providing legal advice, and later sharing the AI-generated materials with counsel could not retroactively make them privileged.  In Fortis Advisors, LLC v. Krafton, Inc., the court similarly treated AI chat history like internet search history for purposes of discoverability.

While these decisions provide guidance, James and Danielle note that “the practical results of the decisions are confusing.”  They explain that although AI prompts and responses are stored with a third-party platform, many conversations occur within closed, password-protected, and sometimes anonymized environments.  Even so, they caution “court decisions to date would caution practitioners who are advising their clients that even closed universe, password-protected prompts and responses are often found to be discoverable.”

James and Danielle also discuss the implications for attorneys, noting that courts are increasingly scrutinizing attorneys' use of AI.  They highlight that some judges have prohibited the use of certain AI tools in their courtrooms, while others have imposed requirements such as maintaining records of AI prompts and certifying the accuracy of AI-assisted filings.  As they explain, “practitioners need to be aware that judges are implementing a wide variety of both restrictive and permissive rules related to the use of AI in their courts that could have a significant impact on a case if not followed.”

The authors conclude that, while the legal landscape surrounding AI continues to evolve and court guidance remains inconsistent, attorneys should take proactive steps to protect client confidentiality, exercise caution when using AI tools, and assume that AI chatbot interactions may be discoverable until the law becomes more settled.