A federal judge in New York has ruled that communications between a defendant and a generative artificial intelligence platform are not protected by attorney-client privilege, a decision that legal experts say establishes a significant precedent for the burgeoning intersection of law and AI. The ruling in United States v. Heppner, delivered by Judge Rakoff of the Southern District of New York, determined that prompts entered by the defendant into an AI system, and the resulting outputs shared with counsel, do not qualify for legal protections.
The case centers on a defendant who utilized a publicly available AI platform – not an enterprise-level, secured system – and then shared the AI-generated content with his attorneys. When the government sought access to these materials during discovery, the defendant asserted attorney-client privilege and work product protection. Judge Rakoff rejected these claims, finding that the communications were not prepared at the direction of counsel and that the AI platform itself is not a confidential intermediary.
The court’s decision hinges on established principles of privilege. According to legal analysis from Husch Blackwell, privilege is “fragile” and extends only to communications made in confidence with the intention of seeking or providing legal advice. The ruling clarifies that simply sharing information *with* an AI system, even if that information relates to a legal matter, does not automatically imbue it with privilege. The court found that because the AI platform’s terms of service allow for data review and potential third-party disclosure, the confidentiality required for privilege does not exist. JD Supra notes that communications with AI, when not directed by counsel, do not constitute attorney work product.
The implications of the Heppner ruling extend beyond active litigation. Freshfields cautions that feeding legal analysis or correspondence with counsel into a public AI system could waive not only attorney-client privilege but also confidentiality and trade secret protections. This is particularly relevant as companies and individuals increasingly turn to AI tools for tasks like summarizing legal documents or brainstorming strategies.
Legal professionals are now advising clients and colleagues to proactively address the risks posed by AI usage. Recommendations include updating client intake procedures to specifically inquire about AI tool usage, incorporating AI-related questions into discovery and depositions, and providing clear guidance to clients about the potential for privilege waiver. Training for regulatory, compliance, and operations staff is also being emphasized to prevent the inadvertent processing of confidential information through public AI platforms. The use of company-approved, closed-universe AI tools is seen as a safer alternative, though even these systems require careful review of their terms of service and user training.
Judge Rakoff acknowledged the novelty of the issues presented, recognizing generative AI as a “new technological frontier” for the law. But, the court underscored that existing legal principles regarding confidentiality and privilege remain applicable. The ruling in Heppner signals a clear message: while AI offers powerful new tools, it does not alter the fundamental rules governing legal privilege.
