Facial recognition is a powerful surveillance tool, one that is deeply unregulated, unjustly biased, and dangerously abused. And yet, over the last few years, agents from the FBI and ICE have conducted thousands of facial recognition searches on immense state driver’s license databases.
According to a report from the Washington Post based on internal documents and emails obtained through public records requests, Utah, Vermont, and Washington have all allowed ICE agents to run facial recognition searches on their state DMV databases. Utah also allowed the FBI to prod its photos. The legal foundation of many of these searches is dubious, and new records again call into question the seemingly unfettered expansion of surveillance and tracking technology. Widely available facial recognition tools without proper safeguards against their abuse could lead to unwarranted invasions of privacy.
Of particular note is that the states covered by the documents grant driving privilege cards for unauthorized immigrants or allow individuals without social security numbers to give an alternate form of documentation.
“The state has told [undocumented immigrants], has encouraged them, to submit that information,” Clare Garvie, a senior associate with Georgetown Law’s Center on Privacy and Technology which obtained the documents, told the Washington Post. “To me, it’s an insane breach of trust to then turn around and allow ICE access to that.”
FBI and ICE agents reportedly ran over 1,000 facial recognition searches in Utah between 2015 and 2017, with dozens of searches a day at times, according to the records. A document from Utah’s Statewide Information & Analysis Center reportedly included guidance on both how to make requests for facial recognition searches and how to take better facial photographs.
While Vermont has banned the DMV from “implement[ing] any procedures or processes for identifying applicants for licenses, learner permits, or nondriver identification cards that involve the use of biometric identifiers” since 2005, the ACLU revealed in 2017 that the government agency had been violating that state law through illegal facial recognition searches. Vermont officials alleged that they had stopped using facial recognition software in 2017, but public records revealed that the agency complied with requests to search their database in the years before then.
In many instances, these requests would be made by emailing the DMV, sometimes with just a description of individuals they were looking for like, “gypsy … scamming elderly people for money” or “VERY LARGE PROTRUDING EARS.” One police officer forwarded their request to the DMV for a man recorded “brazenly” stealing with the text, “Can we play NCIS for this officer?” referencing the fictional television series.
In the instances when a government official would merely email the DMV with the image of their target, an official from the DMV would then run that photo through the state’s driver’s license database and provide any matches to the agent.
Targeted individuals included those suspected of overstaying their visa, giving false information, and stealing, but also extended to witnesses, victims, and other innocent individuals who were not charged or suspected of criminal activity. One search request reportedly identified the targeted individual as part of a “suspicious circumstance.”
There are a number of issues with this casual use of facial recognition technology: It signals the ease at which government officials can access a wealth of state data, how requests for that data are insufficiently regulated, and that the people targeted are members of some of the most vulnerable populations.
What’s more, government officials tap into databases that collect data under the expectation of privacy. When you obtain driving privileges, you don’t consent to having your photo included in a gigantic face recognition database accessed at the whim of an agent with a hunch.
Most insidiously, we still don’t know what the greater implications are of deploying facial recognition searches on a large scale. Many experts argue that law enforcement shouldn’t be legally permitted to use the technology at all. “Building a system with the potential to arbitrarily scan and identify individuals without any criminal suspicion and discover personal information about their location, interests or activities, can and should simply be banned by law,” Andrew Ferguson, a professor at the University of the District of Columbia, said at a congressional hearing in May.
“This power raises questions about our Fourth and First Amendment protections,” Garvie told the Washington Post. “Police can’t secretly fingerprint a crowd of people from across the street. They also can’t walk through that crowd demanding that everybody produce their driver’s license. But they can scan their faces remotely and, in secret, and identify each person thanks to face recognition technology.”
At the congressional hearing in May, Joy Buolamwini, founder of the Algorithmic Justice League succinctly summed up the dangers of this uncontrolled space: “Our faces may well be the final frontier of privacy.”