HeadlinesBriefing favicon HeadlinesBriefing.com

AI Legal Risks: Court Rules Claude Communications Not Privileged

Hacker News •
×

A federal court has ruled that communications with AI platforms like Claude are not protected by attorney-client privilege or work product doctrine. In a February 2026 decision, Judge Jed S. Rakoff found that a defendant's sensitive AI searches could be disclosed to prosecutors, even though they contained legal strategy discussions.

This landmark ruling stems from United States v. Heppner, where the FBI discovered documents showing a defendant's communications with Claude about defending against criminal charges. The court determined that AI platforms are not attorneys, cannot maintain confidentiality due to data retention policies, and that users cannot invoke privilege for communications made without counsel's direction. The decision creates significant legal exposure for businesses using AI.

Commercial real estate faces particular risk as approximately $1 trillion in debt matures in the coming year. With high turnover and pressure to resolve complex deals quickly, professionals might turn to AI for answers, potentially exposing sensitive workout strategies and proprietary information. The court's reasoning suggests that even communications made with counsel's knowledge may not be protected, given AI platforms' data practices.