Legal chats are not legally privileged
I’ll admit: ChatGPT can be a decent stand-in for a lawyer, and so can Claude.
But as OpenAI CEO Sam Altman recently pointed out, conversations with tools like ChatGPT, Claude, or Grok don’t enjoy the same legal privileges that attach to communications with your lawyer, doctor, or therapist. In fact, in most cases, AI chat logs carry no recognized privilege—yet people still share highly sensitive details with these systems.
Both Illinois and New York (and all states to varying degrees) extend special legal protections to information disclosed by a client to their lawyer and by a patient to their doctor or therapist. These protections operate as evidentiary shields that typically prevent compelled disclosure of confidential communications in proceedings like depositions, subpoenas, and trials.
Non-privileged materials, by contrast, are discoverable. Practically speaking, this means a ChatGPT transcript about legal or medical issues will likely be treated like any other non-privileged document—a point Altman conceded in the interview. To make matters worse, voluntarily disclosing attorney-client communications to a third party, like an AI service provider, may even waive privilege as to the original privileged communications.
Self-serving as it may be, Altman’s candor is useful—and it spotlights a policy gap. Until lawmakers define confidentiality or privilege rules for AI-mediated conversations, assuming they ever do, it’s safest to assume anything typed into a chatbot could someday surface in court.
At Replogle Legal Group, we help Illinois and New York businesses negotiate complex deals, build strategic partnerships and protect innovation. Schedule your free consultation today using the online calendar at the link below or contact Ryan Replogle by phone or email.