Can my spouse get my ChatGPT history in a divorce?

Potentially, yes. Chatbot conversations carry no privilege, so a spouse can seek production of chats relevant to finances, custody or settlement, and a party can be ordered to hand them over. In practice the commoner route is simpler still: in one English case a father produced screenshots taken straight from the mother's own ChatGPT account.

Are my AI chats protected like talking to my lawyer?

No. In United States v. Heppner (February 2026) a federal court held that privilege never attached to a defendant's consumer-AI documents: the tool is not an attorney, and the consumer terms defeated any expectation of confidentiality. Separately, work product did not apply because he acted without his lawyers' direction. A consumer subscription fee does not buy legal confidentiality.

Can my employer use my AI chat logs against me?

In litigation, employee and manager AI chat histories can be discoverable evidence. And putting confidential or trade-secret information into a public AI tool can destroy its legal protection: in Trinidad v. OpenAI (January 2026) a trade-secret claim was dismissed with prejudice because the material had been developed through ChatGPT.

Does Secret Chat store chats a divorce or employment lawyer could subpoena from you?

No. Your conversations are stored only in your own browser, not on our servers, so we have no archive to preserve or hand over. The model provider that answers your prompt still processes it under its own terms, and you remain responsible for lawful use — see our Terms and Disclaimers.

Can I use an anonymizer to hide evidence in my case?

No, and you should not try. If you are involved in a dispute, you have a duty to preserve relevant evidence, and destroying or concealing it can be a serious offense. Secret Chat protects lawful, everyday privacy; it is not a tool to evade discovery or the law.

AI Chat Logs in Divorce & Employment Cases

Published July 15, 2026 · Updated August 17, 2026

This article is general information, not legal advice. Laws differ by country and are changing fast. For your own situation, talk to a qualified lawyer — not a chatbot. Your use of Secret Chat is governed by our Terms of Service and Disclaimers, which make clear that you are responsible for your own input and for using the service lawfully.

Two kinds of lawsuits touch almost everyone at some point: the end of a marriage and the end of a job. Both run on evidence — messages, documents, timelines, who-said-what. And in the last two years, a new category of evidence has arrived in both: your conversations with AI. People treat ChatGPT like a confidant while divorcing or job-hunting, not realizing those chats can be pulled into the case. Courts have now shown, repeatedly, that they can be.

Why AI Chats Are Fair Game: No Privilege

Start with the rule that surprises people most. When you talk to your own lawyer, the conversation is protected by attorney-client privilege. When you type the same thing into a chatbot, it generally is not. On February 10, 2026, Judge Jed Rakoff of the Southern District of New York ruled in United States v. Heppner that some 31 defense-strategy documents a criminal defendant had produced using consumer Claude were unprotected — and the precise holding matters far more than the headline. The court did not strip away a protection he had. It held that protection never attached in the first place.

The two doctrines failed for different reasons, and collapsing them into one is the most common way this case gets misreported:

Attorney-client privilege never attached because Claude is not an attorney, and because the consumer terms — which permit collection and disclosure of inputs and outputs — defeated any reasonable expectation of confidentiality.
Work product did not apply because Heppner acted on his own initiative rather than at his lawyers' direction, so the documents embodied no attorney's mental impressions.

The opinion appears to be the first squarely on AI-chat privilege. But read what it actually decided: it concerned the consumer version of one tool under one privacy policy. It did not adjudicate ChatGPT Plus, Claude Pro, enterprise or API products, or tools contractually barred from retaining data — and Rakoff expressly left open that AI use directed by your lawyer might come out differently. So the honest reading is not "every paid tier is worthless." It is that a consumer subscription fee buys you no legal confidentiality, and that no court has held a standalone user–chatbot exchange privileged merely because it concerned a legal problem.

The Other Half: Four Courts That Protected AI Chats

An honest account cannot stop at Heppner, because the months after it went the other way. Four civil courts protected a litigant's AI chats as work product — a separate, more losable doctrine that turns on material prepared for anticipated litigation:

Warner v. Gilbarco (E.D. Mich., February 10, 2026 — the very same day as Heppner): a self-represented plaintiff's ChatGPT queries counted as mental impressions, the court reasoning that generative AI is a tool, not a person, so using it was not disclosure to a third party.
Morgan v. V2X (D. Colo., March 30, 2026): protected, but conditioned — the court ordered disclosure of which AI platforms were used, and kept confidential material out of consumer tools.
Tate Group Automotive v. Legacy Automotive Capital (Business Court of Texas, June 3, 2026): the significant one, because the party was represented by counsel — the very fact Heppner turned on. The court expressly declined to follow it.
Assini v. Hayward (Sup. Ct. Nassau County, June 4, 2026): quashed a subpoena served on the AI provider itself for a self-represented defendant's prompts, uploads and outputs.

So the accurate summary is narrower, and more useful, than "AI chats have no protection." Chats you create for your own case may be protected as work product — in some courts, on some facts, and never automatically. The ordinary conversation you had with a bot months before any lawsuit existed, about your finances or your marriage or your boss, is the one with nothing standing behind it. That is the category most people are actually in, and it is the category this article is about.

For that category the practical consequence is simple: the transcript is just a document. It carries no automatic shield, so it can be sought in discovery like an email or a text message — subject, as always, to relevance, proportionality, and whatever objections a lawyer raises.

Divorce and Family Court: Real Consequences

Family lawyers have moved from warning about this to planning around it. Nothing makes a chatbot conversation privileged in a divorce, so AI chats discussing finances, hidden assets, custody strategy or settlement plans sit in the same category as any other record you hold: a party can be ordered to produce them. To be precise about the state of play — we are not aware of a published divorce ruling compelling production of a spouse's chatbot history. The exposure is structural rather than speculative, and as the English case below shows, it does not need a court order to bite. The everyday scenarios are the damaging ones:

Asking a chatbot "how do I hide money before filing for divorce" — now a written record of intent to conceal assets.
Venting a plan to game custody, or drafting a hostile message you later regret, with the reasoning saved alongside it.
Working out a settlement number or your true financial position, handing the other side your bottom line if it surfaces.

England has already produced the clearest illustration, and it involved no discovery at all. In G v K [2025] EWHC 2961 (Fam), decided on September 29, 2025 by HHJ P. Hopkins KC, a father in a Hague Convention child-return case relied on screenshots of the mother's own ChatGPT prompts, obtained through access to her account: searches about separating from her spouse, a reference to an affair partner, and questions about what evidence she would need to report domestic violence to the police. The judge handled the material cautiously, observing that the domestic-violence searches were open to interpretation in more than one direction, and declined to decide whether the access had been consensual or unauthorised. The lesson is sharper than any discovery rule — the person most likely to read your chatbot history is the one who already has your devices and your passwords.

Employment: When Your Chats Cost You the Case

The workplace side is just as pointed, and one case shows how AI use can actively destroy a legal position. It was not itself an employment suit — it was a self-represented claim brought against OpenAI — but the mechanism transfers directly to any workplace trade-secret dispute. In Trinidad v. OpenAI (N.D. Cal., No. 4:25-cv-06328, January 5, 2026), Judge Jon Tigar dismissed a Defend Trade Secrets Act claim with prejudice because the plaintiff had developed the material she called her trade secret through ChatGPT: having shared it with the tool under its terms, she could not show she had taken the reasonable measures to keep it secret that a trade-secret claim requires. The AI chat did not merely become evidence; it dissolved the very thing she was suing over.

That echoes the incident that first made companies nervous: in 2023, Samsung engineers pasted proprietary source code into ChatGPT to debug it, leaking internal secrets outside the company — after which Samsung temporarily restricted generative AI tools on company-owned devices and internal networks. In any wrongful-termination, discrimination, or trade-secret dispute, an employee's or manager's AI chat history can be discoverable: the message where a manager asked an AI to justify firing someone, or where an employee pasted confidential data, is exactly what the other side wants. (The work-product decisions above are the one real counterweight — and they cover chats created for your own case, not the ones already sitting in your history.)

AI in employment is being litigated at a higher level, too. In Mobley v. Workday, a federal court allowed age-discrimination claims over AI hiring tools to proceed, letting the case reach discovery — a reminder that AI decisions, and the data behind them, are now firmly inside the courtroom.

The Common Thread

Divorce and employment cases look different, but the AI problem is identical in both:

No privilege. AI chats are not protected like talking to your lawyer. Work product is the narrow exception, and only for material prepared for your own case.
Discoverable as ordinary records. Absent that exception they can be requested and produced like any other document you hold — subject to the usual limits of relevance and proportionality.
Third-party disclosure. Typing something into a chatbot can count as telling an outsider — which can waive confidentiality or, as above, destroy a secret.
Stored and retrievable. On most consumer tools the conversation sits on the provider's servers, where it can be preserved or produced — a point we cover in how AI chats can be subpoenaed.

How to Protect Yourself

Take legal strategy to your lawyer, not a chatbot. Only your actual attorney gives you privilege. An AI gives you a transcript.
Assume anything you type could be an exhibit. If you would not want it read aloud in your case, do not type it into consumer AI.
Never paste confidential or company data into a public tool. It can leak, and it can waive trade-secret protection.
Prefer tools that keep no server-side archive for the ordinary, private questions of daily life, so there is no stored history sitting on a company's servers.
If a dispute is already live, preserve, don't destroy. Parties to litigation have their own duty to keep relevant evidence — deleting or hiding it can be a separate, serious problem. No tool changes that obligation; when in doubt, ask your lawyer first.

How Secret Chat AI Fits — and Its Honest Limits

Secret Chat AI is designed to reduce the everyday footprint that these cases feed on, and it is important to be precise about what that does and does not mean.

What it does. Secret Chat keeps no server-side archive of your conversations — your chats and files live only in your own browser (IndexedDB and OPFS), and, in the words of our Terms, "we do not have access to this data, and we do not store it on our servers." It is also an anonymizer: no profile of you, an email used only for account access and payment (never tied to your prompts), and requests that reach the model providers with no name, account, or IP attached. So there is no chat history on our side to preserve or hand over, and a provider-side log is far harder to connect to a specific person.

What it does not do. It is an anonymizer, not a shield against the law and not a content filter. The model provider still has to read your prompt to answer it and may briefly retain it under its own terms; nothing you type becomes immune to legal process; and it does not give you privilege — only a lawyer can do that. Above all, it is not a way to hide or destroy evidence: if you are a party to a case, your own duty to preserve and produce relevant material stands regardless of the tool you use. You are responsible for your own input and for using the service lawfully, as spelled out in our Disclaimers and Terms of Service. The honest, narrow value is real: for the lawful, private thinking of ordinary people, keeping no archive and removing your identity means far less of your life is sitting in a database that a divorce or employment lawyer could one day request.

Frequently Asked Questions

Can my spouse get my ChatGPT history in a divorce?
Potentially, yes. Chatbot conversations carry no privilege, so a spouse can seek production of chats relevant to finances, custody or settlement, and a party can be ordered to hand them over. In practice the commoner route is simpler still: in one English case a father produced screenshots taken straight from the mother's own ChatGPT account.
Are my AI chats protected like talking to my lawyer?
No. In United States v. Heppner (February 2026) a federal court held that privilege never attached to a defendant's consumer-AI documents: the tool is not an attorney, and the consumer terms defeated any expectation of confidentiality. Separately, work product did not apply because he acted without his lawyers' direction. A consumer subscription fee does not buy legal confidentiality.
Can my employer use my AI chat logs against me?
In litigation, employee and manager AI chat histories can be discoverable evidence. And putting confidential or trade-secret information into a public AI tool can destroy its legal protection: in Trinidad v. OpenAI (January 2026) a trade-secret claim was dismissed with prejudice because the material had been developed through ChatGPT.
Does Secret Chat store chats a divorce or employment lawyer could subpoena from you?
No. Your conversations are stored only in your own browser, not on our servers, so we have no archive to preserve or hand over. The model provider that answers your prompt still processes it under its own terms, and you remain responsible for lawful use — see our Terms and Disclaimers.
Can I use an anonymizer to hide evidence in my case?
No, and you should not try. If you are involved in a dispute, you have a duty to preserve relevant evidence, and destroying or concealing it can be a serious offense. Secret Chat protects lawful, everyday privacy; it is not a tool to evade discovery or the law.

Conclusion

The pattern across divorce and employment disputes is now reasonably clear: AI chats carry no privilege, the narrow work-product exception protects only what you prepared for your own case, and on most consumer tools the rest sits in an account on someone else's server. The people hurt worst are usually the ones who treated a chatbot as a private diary during the exact months a lawsuit was brewing. The safest habits are old ones — take legal matters to a lawyer, keep secrets off public tools, and, for ordinary private use, prefer AI that keeps no server-side archive and doesn't tie your words to your name — while remembering that responsibility for lawful use always stays with you.

AI Chat Logs in Divorce and Employment Lawsuits
Real Cases, Real Consequences