More people are asking AI chatbots legal questions before they ever call an attorney. A federal court ruling from early 2026 is a warning about why that carries real risk, especially for anyone facing criminal charges.
What happened in United States v. Heppner
In February 2026, a federal judge in the Southern District of New York ruled on a case involving a financial services executive facing securities fraud charges. The defendant had used the AI tool Claude to help draft and analyze materials related to his defense, then shared those materials with his actual attorneys, later arguing the AI conversations should be protected by attorney-client privilege.
The court disagreed. The ruling found that a consumer AI chatbot is not an attorney, the conversations were not confidential in the way privilege requires, and the exchanges were not made at the direction of counsel. None of the legal protections that normally shield conversations with a real lawyer applied.
Why this matters even though it is not a Florida case
This ruling came out of a federal court in New York, not Florida, so it does not directly control what happens in a Florida courtroom. But the reasoning is being closely watched and discussed by attorneys nationally, including in federal courts here in the Southern and Middle Districts of Florida. The legal principle at issue, that a chatbot is not an attorney and conversations with one are not privileged, is not unique to New York law. It reflects how attorney-client privilege actually works everywhere.
The real meaning of “not privileged”
Attorney-client privilege exists so that a person can speak honestly with their lawyer without fear that those conversations will be used against them. It only applies to communications with a licensed attorney made for the purpose of getting legal advice. A conversation with an AI chatbot does not meet that test. If those conversations become relevant to a criminal case, they can potentially be obtained and used by prosecutors, the same as any other unprotected communication.
Can AI actually represent you in court?
No. Practicing law without a license is prohibited, and that includes AI. No AI tool is a licensed attorney, and none can appear in court on someone’s behalf or form the kind of relationship that makes legal advice legally protected. An AI chatbot can produce confident-sounding answers, but confidence is not the same as accuracy, and an AI cannot be held accountable the way a licensed attorney can.
Where the line actually is
There’s a real difference between reading up on general legal topics and discussing the specific facts of an active or potential criminal case. Once someone is under investigation or facing charges, anything typed into a chatbot about the facts of that case should be treated as something that could later be read by the other side.
If this applies to you
If you are under investigation or facing criminal charges, the safest first step is the same as it has always been: talk to a licensed attorney before discussing the details anywhere else, including with an AI chatbot.
The first consultation is free and confidential, with no obligation to retain. Calling early in a matter usually creates more options than calling late.