You may have utilized ChatGPT, Google Gemini, or another AI tool to conduct some level of research. Or perhaps you described a personal problem on one of these platforms and asked for advice. AI tools are fast, they are available at midnight when you’re anxious and can’t sleep, and they feel private — almost as if your conversations simply disappear.
They don’t disappear. And if you are involved in litigation, sharing information with a third-party AI platform could seriously damage your case.
A federal judge in New York recently made an important ruling. Here is what happened, what it means for you, and what you should do differently starting today.
What Happened in U.S. v. Heppner
In February 2026, Judge Jed Rakoff of the U.S. District Court for the Southern District of New York ruled that 31 documents created by a criminal defendant, Bradley Heppner (“Heppner”), using a third-party AI platform, Claude, were not protected by the attorney-client privilege or the work product doctrine.
Heppner, through his counsel, asserted privilege over these documents, arguing that (1) Heppner transmitted information to Claude, including information that he had learned from counsel; (2) Heppner created the documents for the purpose of speaking with counsel and obtaining legal advice; and (3) Heppner shared these documents with counsel. However, Heppner’s counsel conceded that they did not direct Heppner to create these documents.
The court found that Heppner’s claim of privilege did not satisfy any of the three requirements necessary for the attorney-client privilege to attach:
As to the first requirement that the communications stay between the client and his or her attorney, the court found that Claude is not an attorney. Therefore, the communications did not stay between the client and his attorney.
The second requirement is that the communications were intended to be, and in fact were, kept confidential. The court found that Heppner’s communications with Claude were not confidential, not only because Heppner communicated with a third-party AI platform, but also because Claude’s written privacy policy, agreed to by all users, allows Claude to collect data and reserves the right to disclose such data to third parties.
The third requirement is that the communications were made for the purpose of obtaining or providing legal advice. The court found that these communications were not made for such a purpose because Heppner did not intend to obtain legal advice from Claude.
The court also determined that the documents could not be protected as attorney work product. The work product doctrine protects materials prepared by or at the direction of counsel in anticipation of litigation or trial. The court found that since Heppner prepared the documents on his own volition, he was not acting as his counsel’s agent when he communicated with Claude, and the documents did not disclose his counsel’s strategy. Thus, the work product doctrine did not apply.
Accordingly, while Heppner thought he was assisting his defense by submitting information to an AI platform, he was actually preparing material the Government could use against him.
Why Consumer AI Platforms Aren’t Confidential
When you have a confidential conversation with your attorney to obtain legal advice, those communications are protected from disclosure during litigation. The law is designed to encourage honest, complete conversations between clients and their lawyers. However, this legal protection evaporates when information is shared with third parties — and courts are now treating consumer AI platforms as third parties, not as mere research tools.
Here’s the practical problem: the written privacy policies governing free and consumer-grade AI tools typically allow the platform to (1) collect data from conversations (including inputs and outputs), (2) employ such data to train their models, and (3) disclose such data to third parties. Consequently, when you transmit confidential information regarding your lawsuit to ChatGPT or Claude, you are voluntarily sharing that information with a third party.
What This Means for You as a Client
If you are currently involved, or reasonably expect to be involved, in any civil or criminal matter, you should expect that your interactions with consumer-grade AI platforms will receive the same treatment from opposing counsel as any of your social media accounts, meaning you should assume anything you share will be disclosed during discovery. Requests for production of documents specifically targeting parties’ submissions to AI platforms are already becoming commonplace. If you have entered confidential case information into an AI tool, that conversation may be discoverable.
This concern isn’t hypothetical. If someone is stressed and trying to make sense of a difficult situation, that person may overshare information with an AI platform in hopes of finding peace of mind. Such sensitive information may include an account of an incident where someone was injured, a description of physical injuries, an inquiry regarding expected monetary damages, or a summary of a conversation with an attorney. Every one of those inputs, if submitted to a consumer-grade AI platform, could be discoverable.
What You Should Do
Stop putting case details into consumer AI platforms. This means ChatGPT, Claude, Gemini, and any other free or consumer-grade tool. The convenience is not worth the risk.
Tell your lawyer if you have already used AI for your case. If you’ve used an AI tool to research your case, inform your attorney immediately. He or she needs to understand what information may be out there and plan accordingly. Surprises during discovery are always worse than disclosures managed in advance.
Ask before you use any tool for case-related purposes. There could theoretically be contexts in which your attorney could direct you to use an AI tool for a specific purpose and information transmitted under these circumstances could potentially retain protection from discovery. The court in Heppner stated that attorney-directed use of an AI tool sits in a different legal posture than use on a client’s own volition.
Apply the same caution to other platforms. Emails, text messages and social media posts have always carried discovery risk. Submissions on AI platforms are now in that same category and, arguably, carry a higher risk because users often divulge more information to a chatbot than they would in a text.
The Bottom Line
The law is still catching up to recent developments in AI. Courts across the country are working through legal pitfalls that may result from its use, and the rules will continue to evolve. What Heppner makes clear, at least for now, is that your conversations with consumer AI platforms regarding your legal situation will not receive the same protection from discovery as communications with your lawyers.
If you need legal advice or want to assist in legal strategy regarding your case, call your lawyer. That conversation is protected. One with an AI chatbot is not.
