US Courts Split on AI Chat Privacy in Legal Cases as Israel Faces Similar Challenges

As artificial intelligence (AI) tools like ChatGPT, Claude, and Gemini become widely used for legal advice, courts in the United States are grappling with whether conversations with AI are protected by attorney-client privilege. This issue arises in both civil and criminal cases where parties may be compelled to disclose their AI interactions during discovery processes. For example, in February, a federal court in New York ruled that a former executive under investigation for investor fraud must disclose his AI chats, rejecting claims that these communications were protected by attorney-client privilege since no lawyer was involved and the AI's privacy policy allowed data sharing. Conversely, a Texas business court recently ruled that some AI-generated materials prepared by an employee in a $150 million automotive dispute are protected under work-product privilege, as the employee used AI to organize thoughts rather than to reveal information to the opposing party.

Similar rulings have emerged in Michigan, Colorado, and New York, reflecting a growing judicial divide in the US over AI chat confidentiality. In Israel, courts currently uphold strong protections for materials prepared for trial, but experts warn this may change as AI use becomes more common. Dr. Limor Zar-Gutman from the College of Management notes that using public AI tools without confidentiality guarantees could be seen as waiving privilege. Attorney Efrat Shuster emphasizes that private individuals consulting AI without lawyer involvement face weaker claims to privilege, likening AI chats to internet search histories or app messages that may be subject to disclosure.

Yoram Lichtenstein, co-chair of Israel's national AI committee at the Bar Association, warns that sharing information with AI systems that learn from user input risks waiving confidentiality and increasing exposure in legal discovery. He advises caution and recommends using AI models that do not retain or train on user data. The emerging US legal precedents are expected to influence Israeli courts, potentially requiring disclosure of AI communications in some cases. As AI consultations become routine, legal professionals and clients alike must consider the privacy risks when discussing sensitive legal matters with AI tools.