by Michael Dean Thompson

Through a series of public records requests to the Hennepin County Sheriff’s Office, Unicorn Riot (a non-profit, media-based organization of journalists) has obtained rare insight into how the police department responsible for George Floyd’s death uses technology to spy on its citizens. In all, they gathered documents showing how the police use facial recognition systems, license plate readers, cell phone data, and more.

Among the downtown Minneapolis buildings, the High Intensity Drug Trafficking Area Investigative Support Center (“HIDTA”) and the Real Time Analysis Center (“RTAC”) share rented office space and act as a hub for police tactical operations planning. Together, they serve as a clearinghouse monitoring active field operations in real time and gathering information on the citizens of Minneapolis.

Among the most troubling intelligence gathering systems used by the North Central HIDTA is the facial recognition systems. The technology is notorious for good reason. Among its greatest challenges is that its success rate drops significantly when identifying people of color, especially women of color, leading to frequent misidentification. In addition, the sources used for identification can be social media sites, which are unreliable at best and are likely the result of a warrantless search. Facial recognition additionally threatens our civil liberties and our privacy because, as the American Civil Liberties Union (“ACLU”) says, it allows “governments, companies, and individuals to spy on us wherever we go.” It paints a picture unimaginable even to George Orwell where anonymity is no longer possible, whether in your house of worship or at a protest.

Accordingly, the HIDTA and Criminal Information Sharing and Analysis (“CISA”) unit use facial recognition for “lead generation.” Of the more than 1,600 searches logged in 1,395 days, approximately 679 were tagged with case numbers. Of those searches, many were for property and other nonviolent crimes. Nearly 54 percent — or 921 — of the searches were untagged, however. There are no means available to determine if those searches were for legitimate law enforcement functions.

Among the few changes to the Minneapolis Police Department (“MPD”) in the wake of George Floyd’s death was a ban on using facial recognition. However, even after the ban was passed on February 12, 2021, and as they waited another six days for the mayor to sign it, the cops continued to use the technology. In one case during the lull between the ban’s passing and signing, they failed again to provide any justification for its use.

This is important because in the U.S., 40 percent of police spouses self-report to have been victims of domestic violence in contrast to just 10 percent of the general population. Unfortunately, if the United Kingdom is a representative example, cops are conversely investigated and convicted at much lower rates than the general public. The No Tech for Tyrants November 2022 Report pointed out that “Police officers [in the U.K.] who are abusers are likely to know how to manipulate the system to avoid penalty and/or shift blame to the victim. Furthermore, when officers involved in investigations into domestic violence by their colleagues, they may protect them by covering it up.” Without a valid, logged justification for searching a database, we cannot know if the search was for a murder suspect or the latest love interest of some cop’s daughter.

Paradoxically, more searches were conducted after the ban’s going into effect. As recently as September of 2022, the MPD joined with an external Minnesota agency to perform a facial recognition search because the ban cannot stop the MPD from “complying with the National Child Search Assistance Act ... or other statutes requiring cooperation in the search for missing or exploited children.” It is yet another example of how police departments are adept at finding loopholes to rules that limit the use of controversial and dangerous technologies.

Within Minneapolis’ toolbox is a product that allows for datamining of automotive computer systems, Berla. Cyber Scoop, says Berla, is capable of breaking “into a vehicle’s data systems to find, organize and analyze a vast range that often includes far more than what the car generates itself.” It may not be surprising to discover that modern automotive computing systems store information such as where the car has been, when its doors were opened and closed, how fast it was driven, and even whether seatbelts were used. A recent article by Forbes quoted a Customs and Border Patrol (“CBP”) agent who had written about the search of a 2019 Dodge Charger and how car infotainment systems provide information “used to facilitate the transportation or movement of noncitizens without legal status into and throughout the United States,” including IP and email addresses, phone numbers, and search histories of anyone who had ever brought their phone into the vehicle. The agent claimed the information could indicate the account user’s state of mind, including knowledge, motive, and voluntariness, regarding the offenses under investigation.” It might even be able to reveal passwords.

Not all automobiles create records of the vehicle history and the people within it. Nor is that information always available in real-time. Automated license plate readers fill that void. Minneapolis’ HIDTA uses multiple license plate readers, including from the vendors Vigilant and Clear. Vigilant claims it provides “a nationwide database of license plate reader data via a secure web-based application and allow intelligence agents to analyze the movement of targeted vehicles as they travel the nation’s highways,” according to Department of Homeland Security documents obtained by Vice and quoted by Unicorn Riot. Given a license plate number, HIDTA systems can track cars that pass their cameras and pull up its history with the driver completely unaware.

Cell phones are also a common point of attack for police. With a suspect phone in hand, they can break into it using the Israeli software Cellebrite or other similar tools. Some of the products have their own analytics that can download images, location, search histories, contact lists, and more to create a complete workup on the phone’s owner and any devices to which it might have connected.

Adding to the bird’s eye view of Minneapolis is the increasing reliance on a National Guard RC-26 drone. “DAGGER04” spends hours at a time circling over the Twin Cities. Air traffic control records indicate the DEA, FBI, DHS/CBP, and State Patrol all have their own spy planes and helicopters sharing the view.

The systems described here provide a remarkable amount of information to the police with little to no oversight and rarely a warrant. Cell phone data can be queried from providers on nothing more than a “reasonable suspicion” thanks to Title 28 of the Code of Federal Regulations Part 23. From call records, which can include locations, who called or was called, and for how long, the police can include histories from their various systems to generate a “pattern of life” analysis that can reveal information about people who were never even targeted. And, if the MPD was willing to sidestep a simple ban on using facial recognition systems, how far are they willing to stretch the legality of their remaining tools?

Sources: UnicornRiot.ninja; Criminal Intel Files Show Facial Recognition, Warrantless Surveillance in Minnesota, 19 Jan 23

Subscribe today

Already a subscriber? Login