Recent Decisions Spark Questions on GenAI, Privilege, and Privacy Expectations (ktslaw.com)

Recent federal court decisions have raised crucial questions regarding the intersection of generative AI, attorney-client privilege, and privacy expectations. In the case of *Heppner*, a court ruled that documents created using Anthropic’s AI tool Claude were not protected by attorney-client privilege since they represented communications between the defendant and the AI, rather than the defendant and their counsel. The court emphasized that any documents lacking confidentiality could not acquire privilege simply by being shared with a lawyer. Conversely, in cases like *Warner* and *Morgan*, the courts recognized that pro se litigants could claim work product protection for materials generated with AI platforms, as such disclosures do not equate to sharing sensitive information with an adversary. These rulings highlight the varying interpretations of privacy expectations when using public generative AI tools. While one judge pointed out that users have no substantial privacy rights when communicating with AI platforms, another argued that users might reasonably expect some level of confidentiality similar to that provided by traditional digital communication tools. As generative AI becomes integral to modern workflows, these decisions could shape how legal professionals navigate privilege and privacy, prompting calls for clearer regulations to adapt legal standards to the realities of technology in legal contexts. Legal practitioners are now advised to avoid using public AI tools for privileged work and to engage counsel early to safeguard sensitive communications.