Legal System Adapts as AI Chats Lose Attorney-Client Privilege
By John Nada·Apr 16, 2026·4 min read
A federal court ruling has ruled that AI chat conversations lack legal privilege, prompting major law firms to adjust their strategies accordingly.
A federal court ruling has sent shockwaves through the legal industry, determining that chats with AI tools like Claude lack attorney-client privilege. The decision arose from the case of Bradley Heppner, a fraud defendant who used Anthropic's Claude to strategize his defense. Judge Jed Rakoff's ruling stated that because Claude is not a licensed attorney, communications made through the AI cannot be protected as privileged information.
In response, over a dozen major U.S. law firms have quickly adapted their strategies. They've begun issuing client advisories that warn clients about the risks associated with using AI chatbots for legal discussions. Some firms are embedding these warnings directly into contracts, advising clients that communicating with AI could waive their attorney-client privilege. For instance, New York firm Sher Tremonte has added specific language to its engagement agreements that outlines the potential risks of sharing privileged communications with AI platforms. This proactive approach is believed to be among the first attempts to formally translate a court ruling into a contractual obligation for clients.
Legal professionals are now urging clients to use only 'closed' enterprise-grade AI systems, recognizing that even these tools have yet to undergo thorough judicial scrutiny. The urgency of these adaptations reflects a shift in how AI tools are perceived in legal contexts. Debevoise & Plimpton has provided tactical advice to clients, suggesting that they specify in chatbot prompts that their use of AI is guided by counsel. This guidance aims to invoke the Kovel doctrine, which could extend attorney-client privilege to non-lawyers acting as the attorney's agents. Such strategic maneuvers highlight the legal profession's need to navigate these new challenges with care and foresight.
The ruling marks a watershed moment in AI and attorney-client privilege law in the United States. It's a wake-up call for a legal profession increasingly reliant on AI for client interactions. With clients often turning to chatbots for legal guidance, the implications of these conversations becoming evidence in court are profound. Judge Rakoff's opinion has opened a dialogue on how legal standards will evolve in response to these technologies and the potential pitfalls they present.
The legal landscape remains unsettled. In other cases, such as Warner v. Gilbarco, courts have offered some protections for self-represented litigants who use AI tools, arguing that sharing information with software does not equate to disclosing it to an adversary. This has created a notable distinction in U.S. evidence law: represented parties using consumer AI chatbots are exposed, while self-represented individuals may have more protection. For example, a Colorado court reinforced this logic in Morgan v. V2X, protecting a pro se litigant's AI work product while also imposing conditions on the use of AI tools. The pattern is taking shape: if you're a represented party who decided on your own to use a consumer AI chatbot, you're exposed, whereas self-represented individuals might have a greater shield.
As the legal profession navigates these new challenges, the potential for more court rulings on the use of AI chats as evidence is likely. Legal experts, such as Justin Ellis of MoloLamken, have noted that more rulings will eventually clarify when AI chats can be used as evidence. Until then, the legal profession's version of that clarity is showing up in engagement letters and client communications, with advice that would have seemed strange two years ago: think carefully about what you type into a chatbot, because someone else may read it. The implications of this ruling and the surrounding legal discourse are likely to resonate throughout the legal community as firms attempt to adjust to a landscape where AI tools are increasingly common, but fraught with risks for client confidentiality.
The Los Angeles Superior Court is piloting AI tools for judges, indicating that while the legal field adapts from the client side, the bench is also integrating AI into its workflows. This dual approach highlights the need for the legal industry to establish robust protocols that protect client communications while embracing technological advancements. The integration of AI into judicial processes could offer a glimpse of a future where AI serves as an aide in legal contexts, but it further complicates the landscape for attorney-client privilege.
Law firms must now proactively manage AI interactions to safeguard client rights. The court's decision in Heppner's case not only serves as a legal precedent but also emphasizes the urgent need for law firms to reevaluate their use of technology in client communications. The ongoing conversation about AI's role in the legal field will continue to evolve, as both legal practitioners and clients grapple with the implications of this ruling and the future of attorney-client privilege in an increasingly digital world.