Old Doctrines, New Terrain: the Fourth Amendment, the Stored Communications Act, and Law Enforcement Demands for AI Chat Logs—Key Takeaways From the 2026 IAPP Global Summit

Companies offering generative AI chat technologies to consumers should consider what protections apply under the Fourth Amendment of the U.S. Constitution and the federal Stored Communications Act (SCA) and how they should respond to law enforcement demands for user communications with AI chatbots. 

Those questions, which were the subject of a panel at the 2026 IAPP Global Privacy Summit entitled “The Legal Landscape for Law Enforcement Demands for User AI Prompts and Chats,” are discussed below.

Fourth Amendment 

As consumer use of AI chat technologies skyrockets, whether a user’s communications with an AI chatbot are protected by the Fourth Amendment remains an open question. The panelists asserted that Fourth Amendment protections should extend to a user’s chats under existing precedent, the third-party doctrine exception to the warrant requirement should not apply, and law enforcement therefore must generally obtain a search warrant to access a user’s stored communications with an AI chatbot. 

The panelists grounded their analysis in two recent U.S. Supreme Court opinions addressing Fourth Amendment protections in the digital age: Riley v. California (2014) and Carpenter v. United States (2018). In Riley, the Court held that law enforcement must generally obtain a search warrant for stored cell phone data. Rejecting the government’s argument that a cell phone may be searched incident to arrest, the Court observed that modern cell phones can function as minicomputers capable of revealing “the privacies of life.” Riley v. California, 573 U.S. 373, 403 (2014). Carpenter held that the government’s acquisition of seven or more days of cell-site location information (CSLI) constitutes a search under the Fourth Amendment. Carpenter v. U.S., 585 U.S. 296, 310 n.3, 316 (2018). The Supreme Court in Carpenter rejected the government’s argument that users relinquish any reasonable expectation of privacy in their CSLI when they share those records with their mobile providers, finding that neither rationale underlying the third-party doctrine applied. First, the nature of the CSLI—which reflects an exhaustive, retrospective chronicle of a person’s movements and, in turn, their “familial, political, professional, religious, and sexual associations”—is vastly different and more sensitive than records previously covered by the doctrine, such as bank financial transaction records. Id. at 311-315. Second, an individual’s voluntary exposure of a record to a third party likewise does not support extending the doctrine to CSLI because cell phones—which automatically generate and share CSLI with mobile providers—are indispensable to participation in modern society. Id. at 315. Lower courts have struggled to define and apply Carpenter’s third-party doctrine limitations to different technologies. The Court will have its first opportunity to clarify the scope of that ruling this term when it decides United States v. Chatrie, which asks the Court to decide whether the third-party doctrine applies to stored location data maintained in a Google user’s account. Chatrie will almost certainly affect the application of Fourth Amendment protections to user data generated from emerging digital technologies, including AI chatbots.

The panelists argued that, under Carpenter, users have a reasonable expectation of privacy in their communications with an AI chatbot and the third-party doctrine should not apply. The comprehensive and sensitive nature of such communications, they contended, weighs in favor of a reasonable expectation of privacy because (1) AI chat tools are becoming “indispensable” like cell phones and are used regularly in ways likely to generate comprehensive records of an individual’s thoughts and motivations; and (2) while the CSLI in Carpenter was deeply revealing of the defendant’s religious, political, familial, and romantic associations, user communications with AI chatbots are arguably even more sensitive, as they often contain individuals’ innermost musings on medical issues, philosophical views, and romantic or interpersonal questions. Other Carpenter factors map well onto law enforcement requests for AI chat data: (1) the inherently retrospective nature of a user’s chat history allows law enforcement to develop a detailed catalog of the individual’s internal musings, with such records “subject only to the retention policies” of the provider; and (2) there is relatively little friction for law enforcement to obtain access, because—as with CSLI—a “deep repository” of data can be accessed “[w]ith just the click of a button.” Carpenter, 585 U.S. at 311-312. Further, a user’s reasonable expectation of privacy is not undermined by the fact that queries are submitted to a third-party service provider, and the third-party doctrine should therefore not apply because, according to the panel, conversations with AI chatbots are an integral and essential part of modern life, much like the cell phones in Carpenter.

The panelists also observed that, even where a search warrant is obtained, courts may be more willing to find the Fourth Amendment’s particularity requirement satisfied if demands for AI chat communications are narrowly and precisely tailored, as compared to overly broad keyword warrants listing common query terms or geofence warrants covering large geographic areas. That said, caselaw on keyword and geofence warrants offers a cautionary signal: Law enforcement could draft demands for AI chat communications in a manner that is excessively broad and sweeps in the data of many unrelated consumers. As the panel noted, the Supreme Court’s forthcoming opinion in Chatrie—which is expected to address the propriety of a geofence spanning residential buildings, businesses, and a church—may further illuminate the proper scope of law enforcement demands for digital records, including demands for AI chat communications.

The panelists cautioned that the specific design of an AI chatbot service—including how user communications are collected, stored, and shared—could materially affect the Fourth Amendment analysis. For example, a court could conclude that an individual lacks an objectively reasonable expectation of privacy where the user relies on layered agentic AI technologies or third-party AI plug-ins, given the number of entities that may have access to the underlying data. Such facts, the panel suggested, could make it more difficult to successfully contest legal process.

SCA

The panel briefly discussed whether a user’s stored AI chat communications may be subject to the SCA’s statutory warrant requirement. The panelists agreed that many AI chatbots are likely electronic communications service providers and remote computing service providers as those terms are defined at 18 USC § 2711(2) and 18 USC § 2510(15), and therefore they are subject to the SCA’s various data disclosure restrictions. The panelists also agreed that a user’s stored AI chat communications could reasonably be considered SCA “content”—that is, “information concerning the substance, purport, or meaning of” the user’s communications, 18 U.S.C. § 2510(8)—and therefore be subject to the statute’s warrant requirement. See 18 U.S.C. § 2703(a)-(b). However, they also acknowledged that the SCA’s technical and outdated definitions could persuade a court to conclude otherwise. Furthermore, the SCA, which does not impose a statutory suppression mechanism, has not driven the analysis in cases like Riley, Carpenter, and their ilk.

Conclusion 

While many questions about how the Fourth Amendment and SCA apply to stored user communications with generative AI chatbots remain unresolved, existing case law provides a useful framework for working through them. Significantly, the Supreme Court’s forthcoming opinion in Chatrie may provide valuable framing for a new era of law enforcement data requests directed at AI providers.

For more information on the conference, including several noteworthy panels, please see the following companion blog posts:

• AI Governance—Key Takeaways From the 2026 IAPP Global Summit
• State Privacy Enforcers Discuss Collaboration and Enforcement Priorities: Key Takeaways From IAPP Global Summit 2026
• The California DROP Mechanism—$1.5 Billion in Exposure and the Clock Is Ticking: Key Takeaways from 2026 IAPP Global Summit
• Where Privacy Is Headed Next: Key Takeaways From the 2026 IAPP Global Privacy Summit