Bertillonage in an Age of Surveillance: Fourth Amendment Regulation of Facial Recognition Technologies

The Fourth Amendment is a closed hydraulic system. As a general rule, if government conduct is deemed a “search” under the Fourth Amendment, then agents must secure a warrant from a detached and neutral magistrate before engaging in that conduct. There are exceptions, of course, but they just shift the pressure into another valve. Officers who conduct searches based on their own initiative must show not only probable cause, but also good reasons why, in their circumstances, they were not required to get a warrant. One consequence of these Fourth Amendment hydraulics is a reluctance on the part of the Supreme Court to label government conduct a “search.” As a result, a whole host of government activities that constitute “searches” by any commonsense standard—such as looking for someone in a public place—are immune from Fourth Amendment regulation. Out of fear that officers on the beat cannot be expected to get a warrant before monitoring a suspect’s public movements or demonstrate probable cause in support of these investigative efforts, the Court has left the use of these means and methods to the unfettered discretion of executive agents. Until recently, this state of affairs may well not have threatened “the right of the people to be secure in their persons, houses, papers, and effects against unreasonable searches.” Circumstances have changed with the emergence of new surveillance technologies capable of facilitating programs of broad, indiscriminate surveillance. The Court seemed to recognize this change with its 2018 decision in Carpenter v. United States, which held that government agents need a warrant to access cell site location information. But no matter how salutary, that decision was far from revolutionary. The hydraulics remained intact. By virtue of the nature of cell site location information and its primary investigative role, imposing a warrant requirement did not impose significant burdens on law enforcement or put the technology beyond its realistic, and reasonable grasp. But what about other technologies that are not so well-suited to existing Fourth Amendment hydraulics? Will the Court accept the Hobson’s choice it has created? Or will it chart a new path. This article uses the example of facial recognition technology to argue for a new path. Part I describes facial recognition technology and explains its place in a long tradition of biometric identification. Part II explains the challenges facial recognition technologies pose for the Court’s current Fourth Amendment jurisprudence. Because facial recognition technologies mostly monitor activities conducted in public places and entail neither a physical intrusion nor a violation of reasonable expectations of privacy, their deployment and use does not appear to constitute a “search,” and therefore seems beyond the reach of Fourth Amendment regulation. Relying on both text and history, Part II argues that looking for or trying to find someone using facial recognition technologies constitutes a “search.” This does not mean that government agents must get a warrant before deploying or using facial recognition technologies, which would result in a virtual ban. Rather, as Part III, shows, a more bespoke approach to regulating the deployment and use of these technologies is required. Part III then details a regulatory regime that would allow for the reasonable use of facial recognition technologies while guarding against threats of broad, indiscriminate, and intrusive surveillance