You Talk to ChatGPT Like a Therapist. A Court May Treat the Conversation Like Evidence.

A teenager's ChatGPT history became evidence in a civil lawsuit, and in the 2026 federal case United States v. Heppner, Judge Jed Rakoff ruled that documents a defendant created through Anthropic's Claude were not protected by attorney-client privilege or the work product doctrine, showing that AI conversations lack the legal protections of conversations with lawyers, doctors, or therapists.

You open an AI chatbot and start typing something you wouldn’t put in a public post. Maybe you’re trying to make sense of a problem at work, dealing with a relationship or a decision you don’t know how to make. You explain the situation, add the details and keep going until the chatbot starts to understand what you’re dealing with. It feels like a private conversation but legally, that assumption can get complicated. AI conversations are already showing up in court cases. Sometimes investigators obtain them from an AI company. Sometimes they’re simply found on someone’s phone. And unlike conversations with a lawyer, doctor, or therapist, your chats with an AI assistant generally don’t come with the same kind of legal privilege. That becomes more uncomfortable when you realize this isn’t really a ChatGPT thing anymore. Millions of people now use AI assistants to think through things that are deeply personal. And the more these systems become part of how we work, plan, and make decisions, the more revealing those conversations can become. So what exactly happens when something you thought was a private conversation with AI becomes part of a legal case? Table of Contents Your chat can become part of the case This isn’t a hypothetical privacy concern anymore. The Washington Post reported https://www.washingtonpost.com/technology/2026/08/27/chatgpt-chats-are-being-swept-into-civil-criminal-court-cases/ In one civil lawsuit, a teenager’s ChatGPT history became part of the evidence after lawyers obtained his conversations during the case. The conversations weren’t about committing a crime. They were simply part of how he had used ChatGPT, but they still became relevant to the litigation. A 2026 federal case involving Anthropic’s Claude shows how this can get even more complicated. In United States v. Heppner https://harvardlawreview.org/blog/2026/03/united-states-v-heppner/ , Judge Jed Rakoff ruled that documents a defendant had created through Claude were not protected by attorney-client privilege or the work product doctrine. The defendant had used Claude while preparing for a potential criminal case and later shared the AI-generated material with his lawyer. The important part isn’t that a court has decided AI conversations can never be protected. It hasn’t. The case was decided on its specific facts. What it does show is that using an AI assistant for something deeply personal, or even something connected to a legal matter, doesn’t automatically give that conversation the same protection as a conversation with a lawyer. Once something is written down, it can become part of a record. The fact that the record was created inside a chatbot doesn’t automatically give it a special status. And that leads to the part many people may not realize: privacy and legal privilege are two very different things. Also Read: A Man Tried to Hack a Court AI That Didn’t Exist https://firethering.com/a-man-tried-to-hack-a-court-ai-that-didnt-exist/ Why Private Doesn’t Always Mean Protected Private Doesn’t Always Mean Protected Privacy and legal privilege can sound like the same thing, but they protect very different things. When you talk to a lawyer, attorney-client privilege can protect confidential communications made for the purpose of getting legal advice. Similar protections exist in other professional relationships, such as with doctors or therapists. An AI assistant doesn’t automatically create that same legal relationship. That’s the distinction behind the Heppner case we just looked at. The court wasn’t simply asking whether the defendant’s Claude conversations were private. It was asking whether they met the legal requirements for privilege. Those are two very different questions. You might have a password-protected account, a private chat history, and an expectation that nobody else will read what you wrote. None of that, by itself, creates attorney-client privilege. And that’s easy to overlook because AI assistants increasingly feel like something between a search engine, a notebook, and a conversation with someone who already knows the context. The takeaway is simple: a conversation can feel private without being legally privileged.You can have a conversation that feels completely personal, delete it from your own history, and still not have the kind of legal protection that comes with talking to a licensed professional. This isn’t just a ChatGPT problem The issue isn’t really which chatbot you use. ChatGPT may be the most visible example, but the same questions apply to Claude, Gemini, Copilot, and other AI assistants. Each platform has its own policies, retention rules, account types, and ways of responding to legal requests. So there isn’t one simple rule for AI conversations. What happens to a conversation can depend on the service you’re using, how the account is configured, how long the data is retained and how the conversation is obtained in the first place. And the part here easy to overlook is that the AI company doesn’t always need to hand over your conversation for it to end up in court. Also Read: Small But Powerful AI Models You Can Run Locally on Your System No Cloud Needed https://firethering.com/small-powerful-ai-models-run-locally/ Sometimes the AI company isn’t even involved When people hear that an AI conversation can end up in court, it’s easy to imagine investigators sending a request to OpenAI or Anthropic and getting someone’s chat history back. That can happen, but it isn’t always necessary. In a Missouri case https://www.ozarksfirst.com/news/chatgpt-evidence-springfield-vandalism/ , investigators found ChatGPT conversations directly on the defendant’s phone after he agreed to a search. The AI company didn’t need to provide the messages. They were already sitting on the device. The same basic issue can arise in civil litigation. If someone has access to a person’s device or their account data during discovery, conversations stored there can become part of the material being examined. Once an AI conversation is sitting on your phone, computer, or another synced device, the legal path to that data may have very little to do with the AI company itself. But AI companies can also become part of the chain AI providers can receive legal requests for user data, and their policies generally explain that they may disclose information when legally required. OpenAI, for example, publishes transparency reports https://openai.com/trust-and-transparency/ detailing government requests for user data. There is another, more unusual route too. AI companies can sometimes contact law enforcement themselves when they believe a conversation indicates a serious and credible threat of harm. OpenAI and Anthropic both describe circumstances in which they may disclose information to help prevent serious harm. That doesn’t mean companies are routinely reading everyone’s conversations or handing them over to authorities. The circumstances and legal requirements matter. But it does add another layer to the privacy question. Your AI conversation isn’t necessarily a private record sitting only between you and the model. Depending on the service and the situation, there can be other parties, policies, and legal processes in the chain. What should you actually take away? AI is becoming part of everyday life for a lot of people, so simply avoiding it isn’t really the solution. Being aware of what you’re sharing is. Treat an AI conversation with the same care you’d give any other digital record. If something is highly sensitive, think twice before putting it into a consumer AI account. Check the service’s privacy and retention settings, understand how your conversations are stored, and delete chats you no longer need when appropriate. More importantly, don’t assume that deleting a conversation or keeping it out of public view automatically makes it legally protected. You can think of it like this: if you wouldn’t want those words appearing in a legal file later, be careful about putting them into an AI chat in the first place. For most everyday questions, this changes very little. But as AI becomes a place where people work through increasingly personal and consequential decisions, knowing where that boundary sits is becoming part of using these tools responsibly.