Using AI for Your Personal Injury Case: What's Safe, What's Risky

More and more people turn to AI chatbots like ChatGPT, Claude, Gemini, Copilot, and Grok for advice about things they used to Google, including their own legal problems. If you have been injured and you are thinking about a personal injury claim, it is easy, and increasingly common, to type your questions into a chatbot. You should know that in 2026, doing that with specific details from your case can seriously hurt you.

This guide walks through what recent federal courts have said about AI and legal privilege, the five things you should never type into a chatbot about your case, what you can safely use AI for, and what to do if you have already used AI on your case.

The Heppner Ruling in Plain English

In February 2026, Judge Jed Rakoff of the U.S. District Court for the Southern District of New York issued a decision in United States v. Heppner that matters to anyone thinking about a personal injury claim.

The facts that matter for you: a party had used a consumer AI chatbot to help think about their case. They had fed the chatbot information that, in any other setting, would have been protected by attorney-client privilege. When the other side moved for discovery of those chatbot conversations, the court said the conversations were fair game, and more importantly, that the act of putting privileged information into a consumer AI tool can waive privilege over the underlying communications with the attorney.

The court's reasoning was direct. AI platforms are not law firms. They cannot form an attorney-client relationship with a user. They hold no law license, owe no duty of loyalty, and their privacy policies expressly state that user inputs may be used to train models and may be disclosed to third parties, including government authorities when legally compelled. Given those terms, a user has no reasonable expectation of confidentiality when typing into the chatbot. No reasonable expectation of confidentiality means no privilege.

What Heppner did NOT decide: it did not hold that all AI use by anyone in litigation waives privilege. Attorneys using secure, enterprise-grade tools under professional duties of confidentiality are in a different posture. A second 2026 decision, Warner v. Gilbarco, held that attorney-directed AI work can still qualify as protected work product. The key split is who is doing the AI work and under what terms.

The practical takeaway: if you are the client, you should assume that typing case specifics into any consumer chatbot creates a record that the defense can get to in discovery, and that can damage your case.

Five Things You Should Never Type Into a Chatbot

There are things people have always asked lawyers in private that are now being typed, unprotected, into AI. If a claim is active or foreseeable, keep all of the following out of any consumer chatbot:

1. Facts about the accident itself. How it happened, who was at fault, what you did in the seconds before, what the conditions were. Chatbots often ask follow-up questions that draw out more detail than you intended to share.
2. Emails, texts, or letters from your attorney. Pasting attorney communications into a chatbot is the clearest waiver scenario. Do not do it to "summarize," "simplify," or "get a second opinion."
3. Medical records, diagnoses, or treatment history tied to the claim. These are often the most valuable part of your case. They are also among the most sensitive. Keep them in the doctor-patient and attorney-client channels.
4. Prior statements you gave to insurance adjusters, police, or investigators. Running these through an AI tool creates a parallel record and may invite the defense to argue you have coached or altered your recollection.
5. Anything labeled "confidential," "privileged," or "attorney-client." The labels do not matter legally if you voluntarily share the content with a third party, and a consumer chatbot is a third party.

What You Can Safely Use AI For

AI is not useless for someone with a personal injury claim, it is just risky when used for the wrong things. Lower-risk uses include:

• General legal education. "What is comparative fault?" "What is the statute of limitations for personal injury in my state?" "What is the difference between economic and non-economic damages?" These questions do not disclose facts about your case.
• Organizing notes for your doctor. Typing "help me turn these symptoms into a clean timeline" can help you brief your treating physician. Keep the input focused on symptoms and dates; leave accident-liability details out.
• Finding a personal injury attorney. AI can be a reasonable starting point for questions like "how do contingency fees work?" or "what should I ask at a consultation?"
• Understanding terminology. If your attorney uses an unfamiliar word (deposition, interrogatory, demurrer, MMI, UM/UIM), a chatbot can help you understand the concept without you disclosing any case facts.

A simple rule of thumb: if what you are about to type contains a name, a date, a dollar amount, a diagnosis, or a quote from someone involved in your case, do not type it into the chatbot. If it is an abstract, educational question, you are probably fine.

If You Have Already Used AI on Your Case

Many people will read this guide after they have already used an AI chatbot to talk through their accident. Do not panic, and do not try to fix it quietly. Do these four things, in order:

1. Stop using the chatbot for anything case-related right now.
2. Do not delete the chat history. Deleting records once litigation is foreseeable can look like spoliation of evidence, which is a separate, bigger problem than the original disclosure.
3. Tell your attorney. Do not hide it. Your lawyer cannot protect what they do not know about, and the attorney-client privilege still covers your conversation with your lawyer, including your admission that you used AI.
4. Be prepared for discovery questions. If a lawsuit is filed, you may be asked in interrogatories or at deposition whether you used AI in connection with the claim. Your attorney will coach you on how to answer honestly.

Most AI-related problems are manageable when the attorney learns about them early. They become much harder when they surface for the first time in front of an adjuster or a judge.

How Your Attorney May Use AI

Responsible law firms do use AI, but they use it differently than a consumer does. Expect your personal injury attorney to be using secure, enterprise-grade tools with contractual confidentiality terms, not free public chatbots, for anything tied to your case. These tools typically do not train on client inputs and include data protection terms that support the attorney's professional duty of confidentiality.

In Warner v. Gilbarco (February 2026), a court held that attorney-directed AI work was protected as work product, even when the underlying tool was a generative AI platform, because the work was done under the attorney's professional duties and was not disclosed to an adversary. That is the legal world a competent attorney is operating in.

It is reasonable, and increasingly normal, to ask your attorney about their AI policy during a consultation. Good questions:

• Do you use any AI tools in client matters, and if so, which ones?
• Are those tools enterprise versions with confidentiality protections?
• Do you train model outputs on client data?
• What do you do to protect privilege when using AI?

A firm that cannot answer those clearly is a firm that has not thought about it, and in 2026 that is a yellow flag.

Get Legal Help Before You Type

Personal injury law is fact-specific and state-specific. A chatbot does not know your medical chart, your insurance policy, your state's caps on damages, your venue's jury dynamics, or the defense counsel's tendencies. A real attorney does.

If you have been injured, the best next step is a free consultation with a personal injury attorney licensed in your state. Find a personal injury attorney in our directory.