A recent decision from Judge Jed S. Rakoff of the Southern District of New York marks a watershed moment in the legal profession’s evolving relationship with generative AI. In United States v. Heppner, the court held that a Dallas, Texas, financial services executive's written exchanges with Anthropic’s Claude — used to outline defense strategies, factual narratives, and potential legal arguments — were not protected by either the attorney–client privilege or the work‑product doctrine. United States v. Heppner, No. 25‑cr‑503 (JSR), slip op. at 12 (S.D.N.Y. Feb. 17, 2026).

 

The court found that communications with a public AI platform are not communications with counsel. “Because Claude is not an attorney…that alone disposes of Heppner's claim of privilege.” Id.  Judge Rakoff leaned on “well established” precedent that attorney-client privilege requires three elements: 1) communications between a client and counsel, 2) that are intended to be and are kept confidential, 3) for the purpose of obtaining or providing legal advice. United States v. Mejia, 655 F.3d 126, 132 (2d Cir. 2011). Claude cannot form an attorney–client relationship, and owes no duty of loyalty or confidentiality. Second, the court found that the defendant’s inputs and Claude’s outputs were not confidential, citing the platform’s privacy policy allowing data collection, training uses, and disclosure to third parties, including regulators. Third, the court emphasized sharing output with counsel later that was generated independently and without direction from counsel cannot retroactively make them privileged. Id.

 

The decision also rejected work‑product protection because the AI‑generated reports were not created at counsel’s direction, nor did they reflect the attorney's mental impressions. Instead, they were the product of a client acting alone.  

 

In its ruling, the court rejected a finding of at least two out of the three elements of attorney-client privilege. "In the absence of an attorney-client relationship, the discussion of legal issues between two non-attorneys is not protected by attorney-client privilege." E.g., In re OpenAI, Inc., Copyright Infringement Litig., 802 F. Supp. 3d 688, 699 (S.D.N.Y. 2025).

 

The message is clear: Generative AI is a powerful tool, but using public platforms to explore litigation strategy introduces real privilege risk. Firms must reassess how, when, and where AI tools are deployed — and whether consumer‑grade AI platforms would ever be appropriate for sensitive case development. Thoughtful governance, secure AI environments, and clear client guidance are no longer optional. They are essential.