A U.S. federal court ruled that communications with generative-AI systems are not protected by attorney-client privilege [1, 2].
This decision creates a significant legal vulnerability for business owners and corporate executives who use AI tools for legal guidance. Because these interactions are not confidential, the data provided to AI models could be compelled as evidence in court proceedings [1, 2].
The ruling clarifies that AI tools are not licensed attorneys and therefore cannot establish a privileged relationship with a user [1, 2]. Consequently, any legal problems discussed with systems such as Claude or ChatGPT are not treated as confidential legal discussions [1].
"The court made clear that communications with AI are not protected by attorney-client privilege," John Schroyer said [1].
Legal experts warn that this lack of protection exposes organizations to discovery risks during litigation. When a company uses an AI tool to strategize or analyze legal risks, those prompts and the AI's responses may become accessible to opposing counsel [2].
"When an AI tool becomes a witness, the privilege waiver can expose companies to discovery risks," David A. Hoffman said [2].
The court held that since AI-generated advice does not meet the legal criteria for professional legal representation, the privilege does not extend to these digital interactions [1, 2].
“Communications with generative-AI systems are not protected by attorney-client privilege.”
This ruling establishes a critical boundary between human legal counsel and AI assistance. By denying AI tools the protections of attorney-client privilege, the court is signaling that the legal system does not recognize AI as a professional entity capable of maintaining confidentiality. For corporations, this means that 'shadow AI' use—where employees use unauthorized tools for work tasks—now carries a tangible legal risk of exposing internal strategies and vulnerabilities during the discovery phase of a lawsuit.
