Lawyers advising clients should assume that communications with public AI tools will not be privileged. Credit: SomYuZu/Adobe Stock In the span of six weeks, three federal courts have issued rulings that together begin to sketch the contours of how privilege and work product doctrine will apply to the use of generative artificial intelligence in litigation. The decisions—United States v. Heppner, 2026 WL 436479 (S.D.N.Y. Feb. 17, 2026); Warner v. Gilbarco, Inc., 2026 WL 373043 (E.D. Mich. Feb.