This post is co-authored by CDT intern Emma Li.
The U.S. Court of Appeals for the Fourth Circuit, sitting en banc, recently ruled that the Baltimore Police Department’s Aerial Investigative Research (AIR) program violated the warrant requirement of the Fourth Amendment. The decision is significant because it limits the extent to which law enforcement and other governmental entities will be able to gather and retain individuals’ location information. The ruling is also perhaps the most notable extension to date of the Supreme Court’s reasoning in landmark location privacy case Carpenter v. United States, because it imposes a warrant requirement for location information that is anonymous at the point of collection and obtained by means of aerial surveillance rather than cell tower proximity or GPS coordinates.
In Leaders of a Beautiful Struggle et al. v. Balt. Police Dept. et al, Baltimore community activists sued the Baltimore Police Department (BPD) over its use of several planes to conduct persistent, widescale photographic surveillance. The footage enabled BPD to track the movements of individuals and vehicles near the scenes of serious crimes such as homicides, shootings, armed robberies, and carjackings. An Ohio-based company, Persistent Surveillance Systems (PSS), was contracted to fly the airplanes. They used a “Hawkeye Wide Area Imaging System” that could capture approximately 32 square miles per image per second of surveillance. The cameras photographed approximately 90% of the city for approximately 12 hours a day, and the images were retained for 45 days.
PSS contractors analyzed the surveillance data seven days a week so PSS could report the tracks of vehicles and people at crime scenes, which show where they went to and came from as well as the tracks of the people with whom they met. The images were magnified so that each person or vehicle was visible, but represented only by a dot or blob on a map. Facial geometry, license plate numbers, and vehicle make and model were not visible. On account of public pressure and the opposition of the newly-elected mayor, the program was discontinued, but the Baltimore police retained 14% of the captured imagery data, or 265 hours of coverage, for use in 200 cases.
The BPD argued that the case against it was moot because the surveillance program had not been renewed, and most of the images from it had been disposed of. The Court rejected this argument, noting that 14% of the images had been retained, were being used in pending cases, and could be used in new cases.
The Court ruled that “because the AIR program enables police to deduce from the whole of individuals’ movements . . . accessing its data is a search, and its warrantless operation violates the Fourth Amendment.” The Court applied the Fourth Amendment even though the surveillance at issue was not continuous (it did not occur at nighttime or in bad weather), and even though the “dots on a map” collected in this program were not of identified persons at the time of collection. It noted that locations are “so unique and habitual” that even a few data points can be enough to identify an individual. As a result, particularly when combined with other data that PSS put at the fingertips of Baltimore police, such as license plate reader information, gunshot detection data, and the “CitiWatch camera network,” the information collected was individually identifiable. Indeed, such identification was a purpose of the program.
Two points are particularly notable about the Fourth Circuit’s reasoning. First, the Fourth Circuit stated that Carpenter “applies squarely” to the case, and described how the data collected by the BPD constituted “a retrospective database of everyone’s movements across the city.” It thus extended the Supreme Court’s reasoning in Carpenter beyond the cell site location information at issue in that case to cover other location information that is particularly revealing.
Second, the Fourth Circuit signalled that, when determining whether collection of location information constitutes a Fourth Amendment search, a court should account for the possibility that the location information at issue will be combined with other data to result in surveillance that “surpassess ordinary expectations of law enforcement’s capacity” and “opens ‘an intimate window’ into a person’s associations and activities.” Without saying it was doing so, the Court in Leaders of a Beautiful Struggle applied the “mosaic theory” of the Fourth Amendment, finding that when small data points are combined on a large scale, the Fourth Amendment may require a warrant for the collection as a whole, even if a warrant would not be required for the collection of a single or a few of the same data points.
Prof. Paul Rosenzweig anticipated that courts would take the approach that the Fourth Circuit takes here — grouping non-location information with location information to find that a collection of data points was protected by the Fourth Amendment. This is significant because law enforcement increasingly relies on third-party data brokers to assist in their investigations, including data brokers like PSS that are in the business of grouping different pieces of data to draw inferences. The purchase of certain data from these brokers may require more than money: it may require a warrant under this Fourth Amendment analysis.
The Fourth Circuit was badly split in this case, voting 8-7 to find a Fourth Amendment violation. The primary dissenting opinion argued that Baltimore is a high-crime city and that the majority was taking from the police an important program that was used to combat the city’s wave of homicides. The dissent accused the majority of morphing the ruling in Carpenter “into an effective ban on all short-term warrantless tracking of public movements.” It noted that the Supreme Court had previously green-lighted various forms of aerial surveillance in California v. Ciraolo and Dow Chemical Co. v. United States. Those previous rulings coupled with the need to balance surveillance programs with law enforcement necessities, the dissent argued, meant that the AIR program should have been upheld.
We believe that the majority has the better of this argument. The application of the Fourth Amendment in the digital age requires an assessment of the real-life impact of surveillance. Where the data law enforcement seeks can, in the words of the Carpenter court, “provide an intimate window into a person’s life,” law enforcement officers should be required to obtain a warrant. We know that in today’s world of big data, data aggregation is what provides that intimate window. It is how, for example, internet platforms determine your interests and preferences to target ads. We should not ignore that reality in applying the Fourth Amendment. The fact that law enforcement obtains an aggregation of different types of data to provide a view into a person’s life, rather than using only one type of revealing data, can augment the resulting intrusion into a person’s privacy and make the need for a warrant even more compelling.
Law enforcement officers will argue, forcefully, that the mosaic theory is impossible to apply in practice, because a police officer will not know when the collection of a particular piece of data will “complete the mosaic” and trigger the warrant requirement. While this may be true in the short term, over time, court decisions that embrace the mosaic theory will determine where to draw lines and provide sufficient guidance for law enforcement.
BPD has until November 14, 2021, to submit a petition for certiorari to seek review of this case by the Supreme Court.