Artificial intelligence tools such as ChatGPT, Claude, and other large language models are rapidly becoming part of everyday life. People use them to write emails, summarize documents, and even analyze legal problems.
But a recent federal case — United States v. Heppner — highlights a serious risk: using AI to analyze your legal problems may waive attorney-client privilege.
The decision is one of the first major rulings addressing how traditional privilege rules apply to generative AI. For lawyers, clients, and businesses, the case is an important warning.
The Background of the Case
The case arose in the U.S. District Court for the Southern District of New York before Judge Jed S. Rakoff. The defendant, Bradley Heppner, was charged in connection with an alleged $150 million securities and wire fraud scheme. While the investigation was underway, Heppner used a consumer AI chatbot to analyze his legal exposure and potential defenses.
Heppner entered detailed prompts into the AI system describing the facts of the investigation and asking legal questions. The AI produced written analyses and memoranda discussing possible defenses.
Heppner later shared those AI-generated documents with his lawyers as part of discussions about defense strategy.
When federal agents executed a search warrant and seized Heppner’s devices, prosecutors obtained the AI documents. The defense argued that these materials were protected by: the attorney-client privilege, and the attorney work-product doctrine.
The court disagreed.
The Court’s Ruling
Judge Rakoff ruled that the AI-generated documents were not privileged and could be used by the government. The court identified several reasons.
1. Communications With AI Are Not Attorney-Client Communications
Attorney-client privilege protects confidential communications between a client and a lawyer made for the purpose of obtaining legal advice. An AI chatbot is not a lawyer and cannot form an attorney-client relationship. Because the defendant was communicating with an AI system rather than with counsel, the communications did not satisfy the basic elements of the privilege.
2. Disclosure to a Third Party Destroys Confidentiality
Privilege generally exists only when communications are intended to remain confidential. The court held that entering information into a consumer AI platform was effectively disclosing it to a third party. In this case, the AI platform’s terms of service specifically stated that users should not expect confidentiality in their inputs. Once confidential legal information was shared with the AI system, the confidentiality requirement for privilege was lost.
3. Documents Created With AI Before Talking to Counsel Are Not Privileged
The defense argued that the documents were created to discuss legal strategy with counsel. But the court emphasized a well-established rule: Sending a document to your lawyer later does not make it privileged if it was not privileged when created. Heppner created the AI documents independently and only later transmitted them to his attorneys, which did not retroactively create privilege.
4. The Work-Product Doctrine Also Did Not Apply
The work-product doctrine protects materials prepared by or at the direction of an attorney in anticipation of litigation. The court found that Heppner used the AI tool on his own initiative, not at the direction of counsel. Because the lawyers were not involved in directing the AI work, the doctrine did not apply.
Why This Case Matters
The decision in United States v. Heppner has attracted widespread attention in the legal community because it may be the first ruling where a person’s use of AI resulted in a loss of privilege protection.
It highlights a fundamental principle: Privilege protects relationships — not technology. Attorney-client privilege arises from confidential communications with a lawyer. If a client instead discusses their legal situation with an AI system, that conversation may be discoverable.
Practical Lessons for Clients
The lesson from Heppner is simple but important: Do Not Use AI to Discuss Confidential Legal Issues. If you are facing a legal problem: Do not upload confidential documents into public AI tools. Do not ask AI systems for advice about your legal exposure. Do not summarize conversations with your lawyer in AI prompts.
Those communications may later be obtained by opposing counsel or prosecutors.
Instead, communicate directly with your attorney.
