Hate To Say I Told You So Again: Your Chats Ain’t Private
Above the Law
by Stephen EmbryMarch 2, 2026
AI-Generated Deep Dive Summary
Recent court rulings have highlighted a significant privacy concern in the use of General AI (GenAI) tools like Claude. In *United States v. Bradley Heppner*, Judge Rakoff ruled that communications between Heppner and Claude were not protected under attorney-client privilege or work product privilege, emphasizing that such exchanges lack confidentiality and are discoverable by the government.
The case revolves around Heppner’s use of Claude to strategize his defense, outline legal arguments, and anticipate opposing counsel’s moves. Despite his intent to share this information with his lawyer afterward, the court found no basis for privacy protections due to several factors: the absence of direct attorney involvement, the non-confidential nature of GenAI platforms (which often collect and use data for training), and Heppner’s failure to inform his lawyer in advance.
This ruling underscores critical lessons for legal professionals and clients alike. First, GenAI tools are not a substitute for direct communication with counsel. Second, any information shared with such platforms before consulting a lawyer may be considered discoverable once litigation begins. Additionally, while some AI systems offer privacy options, users must carefully review terms of service to understand data usage policies.
For legal professionals, this serves as a stark reminder to advise clients on the risks of using GenAI tools without proper guidance. Ethical and confidentiality standards under Model Rule 1.6 require lawyers to ensure client communications remain protected, but relying solely on AI for legal strategy bypasses these safeguards. This case highlights the importance of consulting with counsel before inputting sensitive information into any technology platform.
In an
Verticals
legalnews
Originally published on Above the Law on 3/2/2026