Your AI Chat Logs Can Be Used Against You in Court
A federal judge in New York ruled in February 2026 that a former financial services executive must hand over chat logs he created with an AI chatbot, rejecting his argument that the conversations were protected. The decision signals that lawyers and their clients face real legal exposure when using ChatGPT, Claude, Gemini, and other AI tools.
Bradley Heppner, former chair of bankrupt financial services company GWG Holdings, used Anthropic's Claude to prepare reports for his defense against federal fraud charges. Prosecutors demanded access to those chat logs. Judge Jed Rakoff ruled that Heppner must release 31 documents generated by Claude.
The core issue: AI conversations lack attorney-client privilege. Unlike discussions with your lawyer-which are confidential under US law-chats with a bot create no legal protection. Rakoff wrote that an attorney-client relationship "does not exist, or could exist, between an AI user and a platform such as Claude."
Law Firms Are Warning Clients
Over a dozen US law firms have added warnings to their websites. Some, including New York-based Sher Tremonte, now include AI disclaimers in hiring agreements and client contracts.
Alexandria Gutiérrez Swette, a lawyer at Kobre & Kim in New York, put it plainly: "We are telling our clients: You should proceed with caution here."
One firm's client contract stated that sharing a lawyer's advice with a chatbot would likely erase attorney-client privilege entirely. That's a significant risk for clients who use AI tools to supplement legal guidance.
Practical Steps Firms Are Taking
Some firms are exploring "closed" AI systems designed to protect legal communications, though these approaches haven't been tested widely in court yet. Debevoise & Plimpton suggests clients use specific language when seeking AI assistance-for example, prefacing prompts with "I am doing this research at the direction of counsel for X litigation."
Los Angeles-based O'Melveny & Myers and other firms are investigating whether such framing might preserve privilege, but the legal landscape remains unsettled.
What This Means for Your Practice
If you use AI for Legal work, treat those conversations as discoverable. Don't ask a chatbot for legal strategy or advice that you wouldn't want prosecutors or opposing counsel to see.
The Heppner case shows that prosecutors will demand access to chat logs. Courts are granting those requests. Your conversations with AI tools are not confidential.
Consult your firm's AI policy before using any chatbot for client work or sensitive matters. If your firm hasn't issued guidance yet, ask for it.
Your membership also unlocks:
