U.S. v. Heppner, 820 F. Supp. 3d 292 (S.D.N.Y. 2026)
Introduction
As generative artificial intelligence becomes more integrated into everyday life, courts are beginning to confront how existing legal doctrines apply to AI-related conduct. One emerging issue is whether communications between individuals and AI platforms are protected by traditional confidentiality doctrines such as the attorney-client privilege or the work product doctrine. In United States v. Heppner1, the United States District Court for the Southern District of New York addressed this question and concluded that such protections generally do not apply to AI communications.
Factual and Procedural Summary
Defendant Bradley Heppner was charged with multiple counts of fraud, including securities fraud and wire fraud, stemming from his alleged role in defrauding investors of over $150 million. Following his arrest, federal agents seized documents from his home, including approximately thirty-one records of Heppner’s interactions with a generative AI platform known as “Claude.” These communications were created after Heppner became aware he was the target of a criminal investigation and included materials outlining potential defense strategies and legal arguments.
Heppner asserted that these AI-generated materials were protected under attorney-client privilege and the work product doctrine. The government moved to compel disclosure, and the court granted the motion, holding that neither doctrine applied.
Attorney-Client Privilege, Generally
Attorney-client privilege protects confidential communications between a client and an attorney made for the purpose of obtaining or providing legal advice. Courts construe this privilege narrowly because it limits the availability of relevant evidence. To qualify, the communication must:
• Be between a client and an attorney,
• Be intended to remain confidential, and
• Be made for the purpose of legal advice.2
Work Product Doctrine, Generally
The work product doctrine protects materials prepared by or at the direction of counsel in anticipation of litigation. Its purpose is to safeguard the mental impressions and legal strategies of attorneys. However, protection typically applies only when materials are created by attorneys or their agents and reflect counsel’s strategic thinking.3
Reasoning Behind the Decision
The court found that communication between Heppner and the AI platform were not protected because the AI is not an attorney. Therefore, the exchanges were not communications between a client and legal counsel.
Regarding confidentiality, the court also determined that the communications were not confidential. By using a third-party AI platform, Heppner exposed his inputs to potential disclosure, especially given the platform’s privacy policy, which allowed data collection and sharing with third parties. Because Heppner voluntarily shared information with the AI service, he lacked a reasonable expectation of confidentiality. Although Heppner claimed he used AI to assist in preparing to consult with counsel, the court emphasized that the relevant inquiry is whether he sought legal advice from AI itself. AI explicitly disclaimed providing legal advice, and Heppner’s counsel had not directed him to use the platform.
The court held that the work product doctrine did not apply because the materials were not prepared by or at the direction of counsel. Heppner independently created the AI communications, and they did not reflect his attorney’s legal strategy at the time.
Conclusion
The court concluded that communications with a generative AI platform are not protected by attorney-client privilege or the work product doctrine when they are made independently, shared with a third party, and not created at the direction of counsel. As a result, the government was permitted to access and use the AI-generated materials in its prosecution.
Impact and Implications
Heppner represents a significant early decision addressing the intersection of AI and evidentiary privilege. The ruling highlights key risks associated with using AI tools in legal contexts such as (1) sharing sensitive information with AI platforms may waive confidentiality protections; (2) AI systems are not substitutes for legal counsel and do not create privileged relationships; and (3) materials generated without attorney direction are unlikely to receive work product protection.
This case signals that courts will apply traditional legal doctrines strictly, even in the face of emerging technologies, and serves as a cautionary guide for individuals and attorneys using AI in connection with legal matters. The experienced attorneys at Palmer Kazanjian Wohl Hodson are available to advise clients on the legal risks associated with the use of AI. We are glad to offer guidance to clients who want to develop practices that protect confidentiality and strengthen evidentiary compliances.
1 U.S. v. Heppner, 820 F. Supp. 3d 292 (S.D.N.Y. 2026).
2 U.S. v. Heppner, 820 F. Supp. 3d 292, 296 (S.D.N.Y. 2026).
3 U.S. v. Heppner, 820 F. Supp. 3d 292, 297-298 (S.D.N.Y. 2026).