February 17, 2026|Franchise Frontlines

United States v. Heppner: Southern District of New York Holds AI Communications Are Not Protected by Attorney-Client Privilege or Work Product Doctrine

February 17, 2026 | United States District Court for the Southern District of New York | Published Opinion

Executive Summary
In a published decision, Judge Rakoff of the Southern District of New York granted the Government’s motion to compel production of documents created by a criminal defendant through a publicly available generative artificial intelligence platform, holding that such communications were not protected by the attorney-client privilege or the work product doctrine. The defendant argued that the materials reflected legal strategy developed in anticipation of litigation and were later shared with counsel. The Government contended that communications with a third-party AI platform do not satisfy the requirements for privilege or work product protection. The court agreed with the Government, concluding that the communications were neither confidential nor made to an attorney, and that they were not prepared by or at the direction of counsel.

Relevant Background
The case arises out of a federal criminal prosecution alleging securities fraud, wire fraud, and related offenses. Following the defendant’s indictment and arrest, federal agents executed a search warrant and recovered, among other materials, documents reflecting the defendant’s interactions with a generative artificial intelligence platform.

According to the court’s opinion, the defendant had used the AI tool to generate written analyses and materials related to his anticipated defense after becoming aware of the Government’s investigation. The defendant later asserted that these materials were protected from disclosure, arguing that they incorporated information received from counsel, were created in anticipation of litigation, and were intended to assist in obtaining legal advice.

The Government challenged those assertions and sought a ruling that the materials were not privileged and were subject to review. The parties agreed to temporarily segregate the documents pending resolution of the issue.

Decision
The court held that the defendant’s communications with the AI platform were not protected by the attorney-client privilege. In doing so, the court emphasized that the privilege applies only to confidential communications between a client and an attorney for the purpose of obtaining legal advice. The court found that communications with an AI platform do not satisfy that requirement because the platform is not an attorney and does not create the type of fiduciary relationship necessary to support the privilege.

The court also concluded that the communications were not confidential. It relied in part on the AI platform’s publicly available terms and policies, which disclosed that user inputs and outputs may be retained, used for training, and shared with third parties, including governmental authorities. Based on those disclosures, the court found that the defendant could not have had a reasonable expectation of confidentiality when communicating with the platform.

Finally, the court rejected the defendant’s argument that the materials were protected under the work product doctrine. The court explained that the doctrine generally applies to materials prepared by or at the direction of counsel and is intended to protect an attorney’s mental impressions and litigation strategy. Because the defendant created the materials independently, without direction from counsel, and because the materials did not reflect counsel’s strategy at the time they were generated, the doctrine did not apply.

In reaching its conclusions, the court characterized the issue as one of first impression and made clear that the use of emerging technology does not alter the application of longstanding legal principles governing privilege and work product protection.

Looking Forward
This decision is likely to be cited in a wide range of contexts given the rapid adoption of generative artificial intelligence tools across industries. For franchisors and other operators managing distributed systems, the most practical takeaway is not a change in the underlying law, but a clarification of how existing doctrines apply when new tools are introduced into business and legal workflows.

The court’s analysis reinforces that attorney-client privilege and work product protections remain grounded in traditional requirements, including confidentiality, the existence of an attorney-client relationship, and the involvement of counsel in the development of legal strategy. The use of an AI platform—particularly one that is publicly available and governed by standard terms of use—does not, standing alone, satisfy those requirements and may, in some circumstances, undermine them.

From an operational perspective, this case highlights the importance of clearly distinguishing between AI as a productivity tool and legal counsel as a protected advisor. While AI platforms may assist with drafting, organization, or issue-spotting, they should not be treated as a substitute for legal advice. Inputs provided to such platforms may not be confidential, and outputs generated by them may not be protected, even if later shared with counsel.

For franchisors, this has direct implications for how internal investigations, compliance reviews, and dispute-related analyses are conducted. Business teams increasingly rely on technology to evaluate issues quickly, but where legal exposure is implicated, those processes should remain coordinated with counsel to preserve applicable protections. Internal policies and training may be warranted to ensure that personnel understand when the use of AI tools is appropriate and when engagement with counsel is necessary.

More broadly, the decision reflects a consistent approach emerging in early cases addressing artificial intelligence: courts are not creating new, technology-specific legal frameworks, but are instead applying existing doctrines to new factual contexts. The focus remains on the nature of the communication, the relationship between the parties, and the reasonable expectations of confidentiality. In that sense, the legal analysis is unchanged—even as the tools used in business decision-making continue to evolve.


This article is based solely on the opinion of the Court in this matter. The author has not conducted any independent investigation into the facts. For the avoidance of doubt, each statement related to the law and facts in this article is drawn from the Court’s opinion in this case.

Thomas O’Connell is a Partner at Buchalter LLP and Chair of the firm’s Franchise Practice Group. For questions about this article or media inquiries, you can contact Tom at toconnell@buchalter.com.

This communication is not intended to create, and does not create, an attorney-client relationship or any other legal relationship. No statement herein constitutes legal advice, nor should it be relied upon or interpreted as such. This communication is for general informational purposes only and is not a substitute for legal counsel. Readers should not act or refrain from acting based on any information provided without seeking appropriate legal advice specific to their situation. For more information, visit www.buchalter.com.

Practices