AI Chat Risks
A growing number of lawyers across the United States are issuing stark warnings to their clients regarding the use of advanced artificial intelligence
chatbots, such as ChatGPT and Claude. The core concern revolves around the potential for sensitive information discussed with these AI platforms to be exposed and utilized in legal contexts. This advisory has gained significant traction following a US federal court's pronouncement that conversations with AI tools might not qualify for attorney-client privilege. Consequently, these digital exchanges could potentially be brought forth as evidence in both criminal investigations and civil disputes, creating a substantial vulnerability for individuals involved in legal proceedings. Lawyers are emphasizing that the confidential nature of discussions with legal counsel does not automatically extend to interactions with AI, urging a much more cautious approach to protect legal standing and personal liberty.
Court Ruling Impact
The gravity of this situation was underscored by a decision from US District Judge Jed Rakoff in New York. In a notable case involving Bradley Heppner, formerly the chairman of GWG Holdings and facing charges of securities and wire fraud, the judge mandated the disclosure of materials generated by Anthropic's Claude chatbot. Heppner had utilized Claude to assist in preparing his defense, but prosecutors contended that these interactions should not be shielded. Judge Rakoff concurred, explicitly stating that an attorney-client relationship, and by extension its associated privileges, cannot be established with an AI platform like Claude. The court further pointed out that users generally do not possess a reasonable expectation of privacy when engaging with such chatbots. This ruling creates a precarious landscape for clients who might inadvertently compromise their legal protections by confiding in AI tools without fully understanding the implications.
Law Firm Advisories
In response to these emerging legal quandaries, numerous law firms are proactively updating their client communications and contractual agreements. For instance, the firm Sher Tremonte has incorporated language in its client agreements clearly stating that sharing privileged communications with a third-party AI platform could result in the forfeiture of that privilege. This signifies a tangible shift in legal practice, acknowledging the unprecedented challenges posed by AI in maintaining confidentiality. While some legal professionals are advising a complete moratorium on disclosing any information pertinent to ongoing legal matters to AI unless explicitly directed by their attorney, others are exploring alternative strategies. These include the recommendation to utilize 'closed' AI systems designed with enhanced privacy features or to explicitly indicate in prompts that an attorney is supervising the AI's input and output, thereby attempting to introduce a layer of controlled interaction.
Divergent Legal Views
It is important to note that not all judicial interpretations have aligned on this complex issue. In a separate instance, US Magistrate Judge Anthony Patti ruled in favor of a litigant, determining that her conversations with ChatGPT did not need to be surrendered, categorizing them instead as personal work product. Judge Patti articulated that generative AI programs like ChatGPT function as tools rather than sentient entities. This divergence in judicial opinions highlights the evolving nature of legal frameworks in the face of rapid technological advancement. As legal authorities continue to grapple with these questions, further case law is anticipated to provide clearer guidelines on how AI-generated information and related communications will be treated in legal settings. Until such clarity emerges, the prevailing advice from many legal professionals is to exercise extreme caution and prioritize direct communication with legal counsel.
