Procedural History

On January 15, 2026, the United States Court of Appeals for the Ninth Circuit heard oral arguments in United States v. Heppner, No. 25-4172. The case arose from the District of Oregon, where defendant Marcus Heppner moved to suppress AI-generated conversation logs obtained by federal investigators without a warrant.

The district court denied the motion to suppress, finding that Heppner had no reasonable expectation of privacy in conversations conducted with a large language model operated by a third-party commercial provider. The Ninth Circuit reversed.

Facts

In March 2025, federal agents investigating a series of financial fraud schemes obtained records from an AI service provider pursuant to a subpoena. The records included detailed conversation logs between Heppner and the AI system, spanning approximately four months of interactions. The conversations included discussions of financial structures, tax strategies, and — critically — what prosecutors characterized as admissions of fraudulent intent.

Heppner had used the AI system as a planning tool, asking it to analyze various financial arrangements and their legal implications. The government argued that these conversations constituted voluntary disclosures to a third party, invoking the third-party doctrine established in Smith v. Maryland, 442 U.S. 735 (1979).

The Third-Party Doctrine Question

The Ninth Circuit's analysis turned on the nature of AI interactions. Writing for the majority, Judge Chen distinguished between the kind of information at issue in Smith — dialed telephone numbers, which the caller knows are conveyed to the telephone company — and the deeply personal, exploratory nature of AI conversations.

The court noted that users interact with AI systems in a manner more analogous to consulting a professional advisor than to making a phone call. The conversational interface, the system's ability to maintain context across sessions, and the provider's marketing of the service as a private assistant all contributed to a reasonable expectation of privacy.

Analysis

The Heppner decision represents the first federal appellate ruling to directly address Fourth Amendment protections for AI conversation data. Its implications extend well beyond criminal law.

The Privacy Framework

The court adopted a functional analysis that considers several factors in determining whether AI interaction data receives Fourth Amendment protection. These factors include the nature of the interaction, the user's reasonable expectations, the degree of personalization, and whether the data reveals the substance of the user's thought process rather than mere transactional metadata.

This framework echoes the Supreme Court's approach in Carpenter v. United States, 585 U.S. 296 (2018), which rejected a rigid application of the third-party doctrine for cell-site location information. Like Carpenter, Heppner recognizes that technological change can render old categorical rules inadequate.

Implications for the Digital Trust Act

The Heppner decision has direct implications for the Minnesota Digital Trust & Consumer Protection Act's approach to AI agent data. The Act's requirement that AI agents maintain minimal data retention and process information at the edge rather than centrally is consistent with the privacy expectations the Ninth Circuit identified.

If AI conversation data receives heightened Fourth Amendment protection, then systems designed to minimize data retention and maximize user control over their interaction data will be better positioned to comply with both constitutional requirements and statutory mandates.

The Corporate Compliance Dimension

For AI companies, Heppner creates a new compliance obligation. If conversation logs are constitutionally protected, then companies must develop warrant procedures, challenge overbroad government requests, and design systems that can respond to lawful process without compromising the privacy interests the court identified.

This is not merely a criminal procedure issue. The reasoning in Heppner will influence civil litigation involving AI conversation data — in employment disputes, family law proceedings, intellectual property cases, and regulatory investigations.

Conclusion

Heppner is the beginning, not the end, of the judicial conversation about AI and the Fourth Amendment. The Ninth Circuit's framework provides a workable starting point, but the Supreme Court will inevitably need to address the question as the circuits develop potentially conflicting approaches.

The case underscores a broader truth: the legal infrastructure governing AI interactions is being built in real time, through the accumulation of individual cases that each illuminate a different facet of the relationship between humans, AI systems, and the law.

This analysis will be updated as the case progresses through potential en banc review or certiorari.