In United States v. Bradley Heppner, the Southern District of New York ruled that communications with publicly available generative AI do not hold attorney-client or work product privilege. The court determined that there was no attorney-client relationship and no reasonable expectation of confidentiality since the AI’s privacy policy allowed for data sharing with third parties, including government entities. In contrast, the Eastern District of Michigan’s Warner v. Gilbarco, Inc. upheld work product protection for a pro se litigant’s AI-generated materials, noting that generative AI is a tool, not a person. This distinction meant that using AI tools did not waive work product protection, unlike attorney-client privilege, which is waived through disclosure to a third party. These rulings demonstrate that existing legal frameworks apply to AI without creating new doctrines. Understanding these principles is crucial for legal practitioners navigating AI’s impact on attorney-client privilege and work product doctrine.
