Your ChatGPT 'Therapy' Could End Up in Court - Altman Publicly Admits AI Chats Aren’t Privileged

In a recent podcast appearance, OpenAI CEO Sam Altman explicitly confirmed what legal experts have cautioned for over a year: conversations with ChatGPT offer no legal privilege or confidentiality protections – a position that aligns with long ‑ standing advisories from lawyers.

What Did Altman Say?

Speaking on This Past Weekend with Theo Von, Altman addressed how people increasingly treat ChatGPT as a confidential advisor -seeking mental health support, relationship guidance, or legal advice. He emphasized:

“If you talk to a therapist or a lawyer or a doctor… there’s legal privilege… doctor‑patient confidentiality… legal confidentiality. And we haven’t figured that out yet for when you talk to ChatGPT.”

Lawyers Were Already Warned

Legal professionals like Jessee Bundy had urged caution overChatGPT use for legal or sensitive matters. Bundy stated that users treating the AI like a lawyer are effectively sharing discoverable evidence, with no attorney – client privilege, no ethical duty, and no confidentiality protections. Other legal commentators echoed that the claims and interpretations provided by ChatGPT are not privileged communications and could be subject to subpoenas in litigation or regulatory reviews.

Why This Matters

Discoverability
Chat logs – even those deleted by users – might still be stored by OpenAI and could be compelled in court. Currently, OpenAI deletes chats after 30 days unless retention is required by legal or security obligations.

Ongoing Legal Battles
In connection with a lawsuit brought by The New York Times, a court has ordered OpenAI to retain user conversations indefinitely. OpenAI is appealing, arguing this oversteps the company’s privacy commitments.

Privacy Expectations vs. Reality
Many users assume ChatGPT provides a private, protected environment – but legally, it does not. Altman’s remarks shine a spotlight on a growing privacy vulnerability in the public’s use of generative AI

What Users Should Know

Don’t rely on ChatGPT for confidential legal, medical, or therapy advice. It’s not a licensed professional and can’t offer privileged advice.

If you must discuss sensitive content, consider using OpenAI’s enterprise or zero‑data‑retention options – but recognize current legal gaps still apply.

Recognize that anything you say could, in theory, be subpoenaed, unless explicit privilege frameworks are established in legislation or legal precedent – which currently do not exist.

Looking Ahead

Altman voiced the need for what he calls an “AI privilege” – a legal standard akin to doctor‑patient or attorney ‑ client confidentiality for AI interactions. However, creating such a framework would involve navigating difficult questions about professional roles, liability, and the evolving human – AI trust relationship. Courts historically resist expanding evidentiary privilege beyond established domains, making short‑term progress unlikely.

Summary

Sam Altman’s recent comments mark an important turning point – officially affirming what lawyers have warned: ChatGPT interactions are not legally privileged, and user conversations can potentially be leveraged in legal contexts. As the use of generative AI in sensitive areas grows, the urgency for clearer legal and ethical standards around AI privacy becomes undeniable.