Some American lawyers have shared artificial intelligence-related advice for their clients. They cautioned their clients against treating AI chatbots as trusted confidants when their freedom or legal liability is at stake. According to a report by the news agency Reuters, this warning has become more important following a US federal court ruling that said conversations with AI tools like ChatGPT and Claude could be accessed in legal proceedings. Lawyers say such chats may not be protected under the attorney-client privilege and could be used as evidence in both criminal and civil cases.
Legal risk of conversations with AI chatbots: What the ruling means
The concerns follow a ruling by US District Judge Jed Rakoff in New York, who ordered a former executive of a company to hand over documents generated by Anthropic’s Claude chatbot. The case involved Bradley Heppner, former chair of GWG Holdings, who faces securities and wire fraud charges. He had used Claude to prepare materials for his legal defence, but prosecutors argued that these exchanges were not protected, the Reuters report noted.Rakoff agreed, stating that no attorney-client relationship exists “or could exist, between an AI user and a platform such as Claude.” The court also noted that users may not expect privacy in chatbot interactions.Lawyers said this creates risks for clients who rely on AI tools. Alexandria Gutiérrez Swette of Kobre & Kim told Reuters, “We are telling our clients: You should proceed with caution here”. Unlike conversations with lawyers, which are generally confidential, sharing legal details with chatbots could weaken those protections.Several law firms have begun issuing advisories and updating contracts to reflect these concerns. One firm, Sher Tremonte, noted in a client agreement that the “disclosure of privileged communications to a third-party AI platform may constitute a waiver of the attorney-client privilege,” the report added.However, not all courts have taken the same position. In a separate case, US Magistrate Judge Anthony Patti ruled that a litigant did not need to share her ChatGPT conversations, treating them as personal work products. “ChatGPT and other generative AI programs are tools, not persons,” he wrote in the ruling.As legal authorities argue, there is more room for clarification on this front, and more cases will help solidify the treatment of AI-generated information. Until then, some legal professionals advise their clients not to disclose any confidential information related to legal proceedings unless their attorney instructs them otherwise. Meanwhile, other law firms have suggested using “closed” AI tools or incorporating a prompt indicating that a lawyer is supervising the AI use.
