CAIveat Emptor: What You Tell AI Can and Will Be Used Against You (natlawreview.com)

A landmark ruling by Judge Jed S. Rakoff in the Southern District of New York has significant implications for the AI and legal fields, declaring that communications between a defendant and a generative AI platform, Claude, are not protected by attorney-client privilege or the work product doctrine. This decision stems from the case United States v. Heppner, where the defendant’s use of the AI for sensitive legal strategy discussions was deemed non-confidential since the AI does not form a traditional attorney-client relationship, and its data retention policies compromise privacy expectations. As a result, the court allowed the government access to potentially incriminating communications, raising alarms about the risks of using AI in legal and sensitive contexts. For the AI/ML community, this ruling underscores the necessity for businesses, especially in sectors like commercial real estate facing intense pressure, to establish clear AI acceptable use policies. Companies must train employees on safeguarding sensitive information when interacting with AI systems, emphasizing that AI cannot deliver legal advice and lacks the confidentiality required to protect communication. This growing legal clarity demands a cautious approach to AI usage, highlighting the critical importance of confidentiality in the era of AI-driven workflows.