The Fourth Amendment was written to protect against unreasonable searches by government agents who physically rummaged through papers and effects. Two and a half centuries later, the "search" increasingly involves an algorithm scanning thousands of hours of video footage in minutes, detecting faces, reading license plates, and tracking movements across an entire city. The constitutional framework has not caught up, and defense attorneys must be the ones to force the issue.
When law enforcement feeds surveillance footage into an AI system that identifies your client, tracks their movements, or flags their behavior as "suspicious," a foundational question arises: does the AI analysis itself constitute a search within the meaning of the Fourth Amendment? The answer matters enormously, and the courts are only beginning to grapple with it.
The Supreme Court's decision in Carpenter v. United States (2018) marked a watershed moment for digital privacy. Chief Justice Roberts, writing for the majority, held that accessing seven days of historical cell-site location information constituted a search requiring a warrant. The Court recognized that digital surveillance tools have an almost "Orwellian" capacity to track individuals, and that the third-party doctrine could not be mechanically applied to data that reveals "the privacies of life."
Carpenter's reasoning extends naturally to AI-analyzed video evidence. When a police department uses AI to stitch together footage from hundreds of cameras and reconstruct a person's movements over days or weeks, the resulting mosaic of information is functionally identical to the cell-site location data at issue in Carpenter. The individual video clips may each capture activity in public spaces, but the aggregated, AI-synthesized whole reveals far more than the sum of its parts.
Defense attorneys should argue that Carpenter compels warrant requirements for AI-driven video analysis that aggregates footage across time and space. The "mosaic theory," first articulated by Justice Alito in his United States v. Jones (2012) concurrence, provides a framework: even where individual data points are not constitutionally protected, the comprehensive picture created by aggregating those points may be.
Under Katz v. United States (1967), a search occurs when the government intrudes upon a subjective expectation of privacy that society recognizes as reasonable. Courts have generally held that there is no reasonable expectation of privacy in public movements. But AI changes the calculus.
A human officer stationed on a street corner can observe passersby. That is not a search. But an AI system that simultaneously monitors every street corner in a city, identifies every individual, and catalogs their movements in a searchable database is a fundamentally different type of observation. The difference is not just quantitative but qualitative.
Several federal circuit courts have begun to recognize this distinction. In Leaders of a Beautiful Struggle v. Baltimore Police Department (4th Cir. 2021), the en banc Fourth Circuit held that Baltimore's aerial surveillance program, which tracked outdoor movements across the entire city, constituted a Fourth Amendment search. The court relied heavily on Carpenter and the mosaic theory. Defense attorneys should cite this line of authority when challenging AI analysis of municipal surveillance networks.
When confronted with AI-analyzed video evidence, defense attorneys have several avenues for Fourth Amendment challenges.
Motion to Suppress Based on Warrantless AI Analysis. If law enforcement used AI to identify, track, or flag your client without a warrant, move to suppress the results. Argue that the AI analysis constituted a search under Carpenter and Katz, and that no recognized exception to the warrant requirement applies. Be specific about the scope of the AI analysis: how many cameras were searched, over what time period, and what algorithmic techniques were employed.
Motion for Disclosure of AI Methodology. Before you can meaningfully challenge AI evidence, you need to understand what the system actually did. File motions compelling disclosure of the AI system's technical specifications, training data, error rates, and the specific parameters used in the analysis of your case. Prosecutors who resist disclosure are effectively asking the court to admit black-box evidence, and you should frame it exactly that way.
Motion to Exclude Under the Fruit of the Poisonous Tree Doctrine. If the AI analysis led to further investigation that produced additional evidence, the fruit of the poisonous tree doctrine under Wong Sun v. United States (1963) may require suppression of the derivative evidence as well. Map the investigative chain carefully to show that the unconstitutional AI search was the but-for cause of subsequent evidence discovery.
Motion Challenging Exigent Circumstances. Prosecutors may argue that real-time AI analysis of surveillance feeds falls under the exigent circumstances exception. Push back hard. The mere availability of technology that enables faster searches does not create the exigency that justifies bypassing the warrant requirement. If officers had time to set up an AI monitoring system, they had time to obtain a warrant.
The Fourth Amendment requires that warrants "particularly describe the place to be searched, and the persons or things to be seized." When a warrant authorizes AI analysis of video footage, defense attorneys should scrutinize whether it satisfies this requirement. A warrant that authorizes law enforcement to "use facial recognition to search all available surveillance footage in the metropolitan area" is a general warrant of the type the Fourth Amendment was designed to prohibit.
Argue that warrants authorizing AI video analysis must specify the particular footage to be analyzed, the particular individuals or objects to be searched for, and the particular AI techniques to be employed. Anything broader is constitutionally deficient.
There is an important asymmetry here that works in the defense's favor. When the defense analyzes video evidence using on-device tools like FrameCounsel, no Fourth Amendment issue arises. The defense is not a state actor, and the evidence has already been lawfully obtained through discovery. The defense can use AI to analyze every frame of every video without constitutional constraint.
This is precisely why on-device forensic analysis tools matter for the defense. While the prosecution must navigate warrant requirements and constitutional limitations when deploying AI for investigation, the defense can deploy the same analytical power without those constraints. FrameCounsel processes all video locally on your hardware, producing the same caliber of AI-driven analysis that prosecution teams use, but without the constitutional vulnerabilities that make prosecution-side AI evidence challengeable.
The law in this area is evolving rapidly. Several cert petitions raising AI surveillance questions are working their way through the federal courts. State legislatures are beginning to pass statutes regulating AI surveillance, with varying levels of protection. Defense attorneys who develop expertise in these Fourth Amendment arguments now will be positioned to make them effectively as the case law develops.
The Fourth Amendment's protections are only as strong as the lawyers willing to assert them. In the age of AI surveillance, that responsibility falls squarely on the defense bar.
What happens when body camera footage tells a different story than the police report? A guide for criminal defense attorneys on identifying, documenting, and leveraging contradictions between officer narratives and video evidence.
Every time you upload case data to a cloud AI service, you risk destroying attorney-client privilege. Here's why local-only AI is the only safe option for defense work.
How on-device forensic video analysis strengthens admissibility arguments under Daubert, and what defense attorneys need to know about presenting AI-driven evidence in court.
On-device body camera analysis, contradiction detection, and court-ready reports. No credit card required.