Federal Judge Rules AI Chats Lack Privilege

A federal judge in New York has ruled that conversations with artificial intelligence chatbots lack the confidentiality protections afforded to discussions with lawyers, prompting widespread warnings from U.S. law firms to clients facing legal issues. The decision, issued in February, requires a former financial executive accused of fraud to disclose documents generated by an AI tool, highlighting risks as people increasingly seek advice from such systems.

The case centers on Bradley Heppner, former chairman of GWG Holdings, a bankrupt financial services company, and founder of alternative asset firm Beneficent. Federal prosecutors charged Heppner in November 2023 with securities and wire fraud, to which he pleaded not guilty. Before formally engaging defense counsel, Heppner used Anthropic's Claude chatbot to analyze his legal exposure, outline defense strategies and prepare reports, which he later shared with his attorneys. During an FBI search of his home, authorities seized about 31 such documents. Heppner's lawyers sought to withhold them, arguing they contained privileged details related to his defense.

U.S. District Judge Jed Rakoff rejected that claim in the Southern...