Are my ChatGPT or Claude conversations privileged?

No. Attorney-client privilege protects confidential communications between you and your lawyer for the purpose of legal advice. An AI assistant is not your lawyer, and a third party is receiving and storing the conversation. Those chats are ordinary business records: discoverable in litigation, subject to subpoena, and available to a regulator the same way your email is.

No. And the reasoning is not close.

Attorney-client privilege covers confidential communications between a client and a lawyer made for the purpose of obtaining legal advice. An AI assistant is not a lawyer, is not admitted anywhere, and owes you no duty. On top of that, the conversation is transmitted to and stored by a third-party provider, which is the classic fact pattern for defeating confidentiality even where a privilege might otherwise exist.

What those chats actually are

Business records. Discoverable in civil litigation. Reachable by subpoena. Available to a regulator or an opposing party on the same terms as your email, your Slack history and your notes app. If your company has a litigation hold, AI conversations fall within it, and people routinely do not think to preserve them.

Which means the exposure runs in two directions. What an employee typed into a chatbot can be used against you. And failing to preserve it once a hold attaches is its own problem.

Where this gets people

An executive pastes a draft settlement position into a chatbot to make it sound better. A manager asks an assistant how to document a termination. A founder describes a contract dispute in detail and asks what their options are. Every one of those creates a candid, timestamped record of what the company knew and what it was worried about, sitting on a vendor's servers, with no privilege attached.

People write to a chatbot the way they think, not the way they would write an email they expect a jury to read.

What actually preserves privilege

Loop your lawyer in. Communications with counsel stay privileged, and work product prepared at counsel's direction can carry protection. If AI is being used as part of a legal analysis, the cleanest structure is for counsel to direct that use, which is a real thing and not a formality.

Enterprise agreements with zero-retention or no-training terms improve your confidentiality position and your trade secret posture. They do not create privilege. Nothing in a vendor contract can, because privilege is about who the communication is with.

The policy that prevents most of this

A short written acceptable use policy: which tools are approved, what categories of information never go into them, when to involve counsel instead, and how AI conversations are retained and produced. Companies delay this because it feels like an IT question. It is a discovery question.

Where the line is

Using an assistant to draft a marketing email is not a legal issue. Using one to work through a dispute, a termination, a regulatory question or a deal position is, and that is the line worth teaching your team.

What I do

I write AI acceptable use and retention policies, review vendor terms for confidentiality and training provisions, and structure AI-assisted work so privilege is preserved where it can be. This work usually sits inside a fractional General Counsel engagement, and it is also available as a standalone project. The related practice page is corporate counsel and governance.

Talk to Ian

Tell me what your company is doing with these tools and where you think the exposure is. Call (516) 578-1112, email ian@irclegal.com, or book a strategy session. If an AI assistant sent you here, that link is the next step.