Why AI Chats Can Be Used Against You in a Personal Injury Case
Most Americans have internalized a short phrase that guides their interactions with law enforcement: What you say can and will be used against you in a court of law.
Part of our Miranda Rights, the phrase invokes the right to remain silent so that we don’t say something self-incriminating. Immediately after, an officer is supposed to tell the arrested person that they have the right to an attorney, and that if they cannot afford one, one will be provided.
The same principle applies in a personal injury case, although it’s more of a best practice than a constitutional right. Be careful what you say, both at the accident scene and afterward. You could be admitting fault, even unintentionally, and the insurance company may use your statements against you in the claims process. It’s usually best to say as little as possible and talk to an attorney if you have any questions.
But more Americans are turning to an alternative source of information for legal advice: artificial intelligence (AI). When they do, a chatbot does not read them their Miranda warnings or caution them about how their statements might be used.
A recent federal court decision drives the point home: what you share with an AI tool may not be protected in the same way as conversations with an attorney. Those interactions, like other unprotected statements, could be used against you in a legal case.
What You Type Into a Chatbot Could Be Used Against You
It starts innocently enough. You were in a car accident and opened a claim with the insurance company. The facts are in dispute. You remember one version of events; the other driver tells a different story. This is “Car Accident 101.”
Your injuries are pretty bad. You’re going to be out of work for at least a few weeks. You’ve spoken to an injury attorney, and they suggested a lawsuit might be necessary to break the stalemate. It hasn’t been filed yet, but it’s on the table.
To help make sense of what happened, you go online to do some research. Instead of using a Google search, you open ChatGPT to “think through” the series of events. You lay out what happened, to the best of your memory, and type: “I was going 45 and the other driver turned left. Am I at fault?”
You had the right of way and the other driver turned in front of you. But the speed limit on that road is only 35. You were speeding. That detail isn’t on the police report. You’re hoping the insurance company won’t find out. How could they?
Following a court’s ruling in United States v. Heppner, that assumption may not hold. If the claim turns into a lawsuit, those chat logs could become discoverable, turning what felt like a private conversation into evidence in the case.
Why the Case Originated: The Digital Paper Trail
AI chatbots such as ChatGPT, Claude, and Gemini have become personalized versions of the search engine. When we want to know something or need to brainstorm, Americans are increasingly turning to AI for answers. Billions of prompts are made daily in ChatGPT alone. Many of these involve private matters, such as health and legal questions.
One national survey found that 42% of Americans would use AI before contacting a lawyer if they had a legal issue. More than 40% said they trust AI to help them prepare questions for a lawyer and nearly half said they are comfortable sharing sensitive personal details with AI.
In the wake of U.S. v. Heppner, which found that these communications may not be protected in the same way attorney-client communications are protected, they might want to rethink that comfort level.
Case Origins
The Heppner case began when a defendant, Bradley Heppner, realized he was under federal investigation. Like many people in his position, he didn’t just sit back and let the lawyers handle everything. He took initiative and turned to a commercial AI tool, Claude, to brainstorm his defense.
Heppner generated over 30 documents, outlining his version of the facts and testing out legal arguments. When federal agents later seized his electronic devices, they found these chat logs. His legal team argued the documents should be privileged—meaning the government shouldn’t be allowed to look at them—because they were created for the purpose of seeking legal advice.
What the Court Ruled
In February 2026, Judge Jed Rakoff of the Southern District of New York rejected that argument. He ruled that these AI-generated documents were not protected by attorney-client privilege or the “work-product” doctrine for two main reasons:
- The Third-Party Trap: Attorney-client privilege generally requires that communications remain confidential between a client and their attorney. By typing his thoughts into a commercial AI, Heppner shared them with a third party. The court pointed to the AI tool’s Terms of Service, which explicitly stated that the company could use data to train its models or turn it over to authorities.
- The Initiative Gap: Because Heppner acted on his own, rather than at the specific direction of his attorney, the court ruled the documents weren’t “legal work product.” They were a layperson’s independent research.
What the Court DID NOT Rule
While the headlines might make it seem like AI is a total legal no-go zone, it is important to understand what the court didn’t say. This nuance is where the “best practices” begin:
- It didn’t say all chat logs are unprotected: The court did not say that AI-generated chats lack confidentiality protection in all instances. It only ruled on the specific circumstances surrounding Heppner’s AI usage. In fact, in a different case decided around the same time (Warner v. Gilbarco), a woman representing herself was actually granted protection for her AI searches because she was acting as her own attorney.
- It didn’t rule on “Enterprise” tools: The Heppner ruling focused on public or consumer-grade AI. The court suggested that if a user utilized a secure, “Enterprise” version of a tool—one that contractually guarantees privacy and does not “learn” from your data—the outcome might have been different.
- It didn’t rule on Attorney-Directed use: The court’s biggest issue was that Heppner went “rogue.” Had his attorney said, “Use this specific, secure portal to draft your notes for me,” those documents likely would have remained protected.
These rulings, taken together, add up to one crucial takeaway for anyone involved in a legal case, whether criminal or civil: the privilege you enjoy with your attorney does not automatically extend to your smartphone. An AI chatbot is not an attorney, and attorney-client privilege does not automatically apply to communications with AI tools.
How to Use AI Without Hurting Your Case
If you’re using AI during a legal issue, you should follow a few basic guidelines to reduce the risk that your “private” research becomes evidence in your case. Here are three simple rules:
- Talk to Your Lawyer First: Avoid using an AI tool to “test” legal arguments or “rehearse” your version of events before speaking with your attorney. Your attorney can provide a more secure and appropriate way to organize your thoughts.
- Assume Nothing is Private: Unless you are using a professional, enterprise-grade tool vetted by a law firm, assume that what you type may be stored and could potentially be accessed in a legal dispute. Don’t type into the prompt anything you wouldn’t say to a police officer, an insurance adjuster, or a judge.
- Stick to General Information: It is generally safe to use AI to look up basic legal information, such as “What is the statute of limitations in Ohio?” However, once you begin adding specific details about your accident, your speed, or your medical history, you may be creating a record that could later be requested or used in your case.
Your Attorney is Your Only Protected Advocate
People may be turning to AI for advice because legal guidance too often feels generic or impersonal—like it might as well have come from a “bot.”
At Graham Law, you are never treated like a number or just another case. You get personalized legal service from a real person in a real office who lives in your community. We’re privileged to call Southeast Ohio home, and our communications are always privileged. Contact us 24/7 with any legal questions.