Can Your Employer Obtain Your ChatGPT Searches in an Employment Case?

What employees should know before using ChatGPT, Claude, Gemini, or other AI tools to analyze a workplace dispute, potential lawsuit, or employment case.

Reviewed for legal accuracy by William K. Phillips, Founder and Managing Partner, Phillips & Associates, PLLC. Published September 2026. Last reviewed September 2026.

You are being sexually harassed, discriminated against, or retaliated against at work. You open ChatGPT, Claude, Gemini, or another artificial intelligence tool and type out everything that happened. You upload your timeline, your employer's emails, your HR complaint, or even a communication from your lawyer. Then you ask the AI whether you have a case, what your case may be worth, what evidence matters, how to respond to your employer, or what you should do next.

You may have just created evidence that your employer will try to obtain in litigation.

An employer is not automatically entitled to every AI conversation, prompt, or search you have ever made. But you should not assume that a conversation with an AI chatbot is private, confidential, protected by the attorney-client privilege, or beyond the reach of discovery.

Can Your Employer Get Your ChatGPT History in an Employment Lawsuit?

Potentially. An employer may request relevant ChatGPT prompts, conversations, chat histories, uploads, and outputs during discovery in an employment lawsuit. Whether those materials must actually be produced depends on the facts, including their relevance, how the AI was used, whether litigation was anticipated, whether an attorney directed the work, and whether attorney-client privilege or work-product protection applies.

Courts are beginning to reach different conclusions about these issues. Employees should not assume that case-related AI conversations are either automatically protected or automatically discoverable.

Key Takeaways

• ChatGPT and other public AI platforms are not your lawyer.
• AI conversations about an employment dispute may be requested in discovery, but they are not automatically discoverable. Courts are reaching different conclusions about when AI-assisted materials may be protected.
• Uploading or sharing communications with your lawyer through a public AI platform may put attorney-client privilege at risk.
• AI cannot reliably value a sexual harassment, discrimination, retaliation, or other employment case.
• Do not delete existing AI conversations after litigation is reasonably anticipated.
• Tell your employment lawyer about any case-related AI use.

AI Is Not Your Lawyer, and Your AI Conversation May Not Be Confidential

The attorney-client privilege protects certain confidential communications between a client and a lawyer made for the purpose of obtaining or providing legal advice. ChatGPT and other public AI tools are not your lawyer. Typing facts into a chatbot does not create an attorney-client relationship.

The risk grows when you use AI to discuss an existing or anticipated legal claim. The prompts may reveal your version of events, identify witnesses, expose perceived weaknesses, record changing explanations, contain statements that can be compared against later testimony, and show which documents you considered important.

If you upload a lawyer's email, summarize legal advice, or ask the chatbot to evaluate your counsel's strategy, the employer may argue that you disclosed privileged information to a third party and waived the privilege.

What Did the New York Federal Court Decide About AI Conversations?

In United States v. Heppner, No. 25 Cr. 503 (JSR), 2026 WL 436479 (S.D.N.Y. Feb. 17, 2026), Judge Jed S. Rakoff addressed 31 documents reflecting a criminal defendant's communications with Claude, Anthropic's generative AI platform. The defendant argued that the documents concerned his defense, included information learned from his lawyers, were prepared in anticipation of an indictment, and were later shared with counsel.

The court held that the documents were not protected by the attorney-client privilege or the work-product doctrine.

The court's reasoning was direct. Claude was not an attorney. The communications were made to a third-party platform and were not confidential under the platform's privacy terms. And the materials were not prepared by, or at the direction of, counsel. The court further concluded that even if the defendant had entered privileged information into Claude, disclosing it to Claude and Anthropic waived the privilege just as disclosure to another third party could.

Heppner was a criminal case, not an employment case. But the privilege and confidentiality principles the court applied can arise in civil employment litigation as well. An employer can request relevant documents and electronically stored information during discovery. Depending on the claims, defenses, and governing law, that request may include AI prompts, uploads, outputs, chat histories, or documents created with AI.

Are All ChatGPT Conversations Discoverable in an Employment Case?

No. There is no rule that every ChatGPT conversation or other AI interaction becomes discoverable simply because an employee brings an employment lawsuit. Courts are beginning to distinguish between AI materials that may be discoverable and AI-assisted work that may qualify for work-product protection.

Although Heppner found the AI communications at issue unprotected, that decision does not establish that every employer is entitled to every AI conversation.

• In Warner v. Gilbarco, Inc., No. 2:24-cv-12333 (E.D. Mich. Feb. 10, 2026), an employment discrimination case, the court rejected the employer's attempt to compel a self-represented plaintiff to produce all documents concerning her use of ChatGPT in connection with the lawsuit. It treated the AI-assisted materials as protected work product.
• In Morgan v. V2X, Inc., No. 1:25-cv-01991 (D. Colo. Mar. 30, 2026), another employment discrimination case, the court concluded that a self-represented litigant could assert work-product protection for AI use. Using AI to prepare for litigation, the court reasoned, can resemble confidential, strategy-based litigation preparation. But it also required disclosure of the identity of AI platforms used with information designated confidential under a protective order, and imposed restrictions on uploading confidential discovery material to AI platforms that lacked specified safeguards.

For New York employees, the most relevant recent state-court decision is Assini v. Hayward, 2026 NY Slip Op 26086 (Sup. Ct., Nassau County June 4, 2026). There, a New York court considered an amended subpoena directed to OpenAI seeking ChatGPT prompts, inputs, uploaded materials, and corresponding outputs related to litigation filings, claims, defenses, motions, and correspondence. The court considered Heppner, Morgan, and Warner, found the reasoning in Morgan persuasive, and quashed the subpoenas to OpenAI.

Assini was a New York civil case, though not an employment discrimination case. Together, Heppner, Warner, Morgan, and Assini show why employees should not assume there is a universal rule governing AI conversations. Whether AI material is protected or discoverable can depend on who used the AI, why, whether litigation was anticipated, whether a lawyer directed the work, what information was disclosed to the platform, the governing law, and the specific material being requested.

The law is developing. That uncertainty is a reason for caution, not a reason to assume that all AI conversations are protected or that all are discoverable.

What Could Your Employer Learn From Your AI History?

• Your complete factual narrative. The prompt may contain details you never included in your HR complaint, agency charge, demand letter, or deposition testimony.
• Inconsistencies. Small differences in dates, wording, witnesses, or the sequence of events may be used to attack your credibility.
• Your litigation strategy. An AI conversation may reveal the claims you considered, the evidence you believe is missing, your settlement concerns, or what you think the employer will argue.
• Confidential documents. Uploads may include personnel records, internal messages, medical information, trade secrets, or communications with counsel.
• Your damages and mitigation efforts. Prompts may discuss emotional distress, medical treatment, compensation, job searches, or reasons for rejecting other work.

Can AI Accurately Tell You What Your Employment Case Is Worth?

No. ChatGPT or another public AI tool cannot reliably determine what an individual sexual harassment, discrimination, retaliation, or other employment case is worth.

Many employees ask, "How much is my sexual harassment or employment discrimination case worth?" The tool may respond with a specific dollar range and sound completely confident. That confidence does not make the valuation accurate.

AI does not have access to the information needed to value an individual employment case. Many sexual harassment, discrimination, and retaliation matters resolve privately, including before a lawsuit is filed, and many settlements are confidential. Those results are not collected in a complete public database that an AI tool can reliably search. Public verdicts and reported settlements represent only a limited and potentially distorted portion of actual employment-case resolutions.

There is also no reliable national statistic establishing that 98% of sexual harassment or employment discrimination cases settle. One major study of federal employment discrimination cases reported that almost 70% were terminated by settlement and that 7.88% reached the trial stage, and the researchers concluded that no reasonable estimate supported the frequently repeated claim that more than 90% of all filed federal cases settle. Those figures do not include the unknown number of claims resolved confidentially before filing, and they do not tell the public what employers paid in confidential settlements.

AI is generally working from published verdicts, publicly announced settlements, general internet articles, and whatever facts you chose to include in the prompt. It cannot assess missing documents, witness credibility, the employer's evidence, how the decision-maker will testify, the strength of likely defenses, the applicable damages law, mitigation, venue, arbitration risk, the employer's settlement history, or how experienced counsel on both sides will evaluate the case.

Do not make a settlement decision, reject an offer, or set your expectations based on a number generated by AI. It may rest on incomplete facts and a tiny, unrepresentative portion of the cases that actually resolve.

A real valuation requires legal analysis, experience, and business judgment. An experienced employment lawyer, not a chatbot, is the person to evaluate what an employment case may be worth: liability, economic damages, emotional distress, available evidence, corroborating witnesses, employer knowledge, retaliation, mitigation, venue, litigation risk, likely defenses, and the practical ability to prove each claim. That value can change as documents are produced, witnesses are interviewed or deposed, and the employer's explanation develops.

Why We're Qualified to Evaluate Employment Cases

We represent employees exclusively and practice employment law exclusively. We combine the resources of a substantial plaintiff-side employment practice with partner-led, boutique-style representation. We have handled more than 9,500 employment matters, litigated approximately 2,000 employment cases in federal and state courts, and recovered more than $360 million for employees. Prior results do not guarantee a similar outcome.

A Proprietary Case-Valuation Database Built From Actual Employment Cases

We use a proprietary case-valuation database designed and developed by our Founder and Managing Partner, William K. Phillips, a former investment banker. It tracks our resolutions together with factors including the underlying facts, claims, employer, defense counsel, courthouse, judge, and outcome. That lets our attorneys compare a current matter against actual cases we have handled, including confidential resolutions that are not available in public verdict databases or ordinary internet searches.

Case valuation here is not one lawyer's opinion or a single comparable result. Multiple partners and attorneys evaluate significant matters using our case data, institutional experience, and litigation history, and the potential value, risk of dismissal, cost and burden of continued litigation, and benefits and risks of settlement are reassessed throughout the matter as the evidence and defenses develop.

Partner-Led Litigation Teams

We have 38 full-time, in-house attorneys organized into 18 dedicated litigation teams, with more than 200 years of combined employment law experience. Every accepted matter is assigned to a dedicated legal team led by a partner or senior litigator. The same team can develop the facts, evaluate damages, negotiate, conduct discovery, take and defend depositions, brief motions, and prepare the case for trial and appeal.

Experience Before Federal and State Courts

Our attorneys have appeared before more than 110 United States District Judges and more than 70 United States Magistrate Judges. That experience helps us evaluate venue, judicial history, motion practice, discovery risk, trial exposure, and how the same evidence may be received at different stages of a case.

Experience Opposing Management-Side Defense Firms

We have litigated against more than 550 distinct management-side defense firms, including Jackson Lewis, Littler Mendelson, Ogletree Deakins, Seyfarth Shaw, Morgan Lewis, and Proskauer Rose. Employers, insurers, and defense lawyers evaluate whether an employee's attorneys are prepared to continue through discovery, depositions, motion practice, trial, and appeal if a case does not resolve. That litigation record provides practical context a public AI chatbot cannot replicate.

You can review representative case results from our work.

What Should You Do If You Used AI to Discuss Your Employment Case?

• Do not use a public AI chatbot to analyze the facts, documents, value, or strategy of your legal case.
• Do not paste, upload, summarize, or ask AI to rewrite communications with your lawyer.
• Do not upload workplace documents without first getting legal advice about whether you may lawfully retain or disclose them.
• Do not use an employer-owned device, network, email address, browser profile, or AI account for personal legal research. Separate workplace-monitoring rules may give the employer access.
• If you already used AI, tell your lawyer exactly what you used, what you entered or uploaded, and what the tool produced.
• Do not delete chats, prompts, outputs, or uploaded files after a dispute or lawsuit is reasonably anticipated. Deleting potentially relevant material can create a separate evidence-preservation problem.

The Bottom Line

AI can feel private because the conversation takes place on a personal screen. Legally, that is a dangerous assumption.

If you are involved in a career-impacting employment dispute, treat everything you enter into a public AI platform as something your employer's lawyers may later request, review, and use. Ask your attorney, not a chatbot, for legal advice about your case.

We represent employees in sexual harassment, discrimination, retaliation, and other employment matters. If you believe your workplace rights have been violated, call (866) 229-9441 or contact us for a confidential consultation before using AI to analyze or organize your potential claims.