The US government is deploying a new facial recognition system at the southern border that would record images of people inside vehicles entering and leaving the country.
The pilot program, scheduled to begin in August, will build on secretive tests conducted in Arizona and Texas during which authorities collected a “massive amount of data”, including images captured “as people were leaving work, picking up children from school, and carrying out other daily routines”, according to government records.
The project, which US Customs and Border Protection (CBP) confirmed to the Guardian on Tuesday, sparked immediate criticisms from civil liberties advocates who said there were a host of privacy and constitutional concerns with an overly broad surveillance system relying on questionable technology.
How big is the force?
Already the largest and most funded federal law enforcement agency in its own right, the border patrol is part of the umbrella agency US Customs and Border Protection (CBP). CBP’s approximately 60,000 employees are split in four major divisions: officers who inspect imports; an air and marine division; agents who staff ports of entry – international airports, seaports and land crossings; and the approximately 20,000 agents of the border patrol, who are concentrated in the south-west, but stationed nationwide.
What are its powers?
The border patrol enjoys extraordinary police powers. Agents operate checkpoints where they stop motorists everyday without suspicion, and in the interior of the country up to 100 miles, they can board planes, trains and buses. There is no geographic limit to which agents can otherwise conduct stops amid roving patrols, though they are technically required to have reasonable suspicion to do so. Extending from not only land borders but also the Pacific, Atlantic, Gulf and Great Lakes coasts, the 100-mile zone encompasses two out of every three Americans, 12 states in their whole or near entirety and nine of the 10 biggest cities in the nation.
“This is an example of the growing trend of authoritarian use of technology to track and stalk immigrant communities,” said Malkia Cyril, the executive director of the Center for Media Justice. “It’s absolutely a violation of our democratic rights, and we are definitely going to fight back.”
The so-called Vehicle Face System, first reported by the Verge, will run for one year at the Anzalduas port of entry in Texas, tracking cars traveling to and from Mexico with the goal of testing the camera’s “ability to capture a quality facial image for each occupant position in the vehicle” and the “biometric matching accuracy” of the images, CBP said. Authorities will “compare” those images with ones stored in “government holdings”, which include passports, visas and other CBP documents, said a spokeswoman, Jennifer Gabris.
Last year, Oak Ridge National Laboratories, a government-sponsored lab, obtained approval to test the new cameras, determining that the technology was “capable of capturing a high quality image” of drivers’ faces and possibly other occupants in moving cars.
The US has been aggressively expanding its monitoring and targeting of people at the southern border and surrounding regions, and there have been increasing concerns about border agents and the Transportation Security Administration (TSA) searching electronic devices.
But this appears to be the first time CBP will be utilizing this kind of facial recognition technology that records images from cars at the border, said Mitra Ebadolahi, a staff attorney with the ACLU’s Border Litigation Project.
“Once these kinds of powerful surveillance systems are built and deployed, the privacy harms … can’t be undone,” said Ebadolahi. “People don’t understand just how invasive these technologies are, and people don’t even know they are being targeted.”
Research has repeatedly shown that facial recognition is a “fundamentally biased technology”, she added.
Critics have increasingly warned that biometric tools can exacerbate existing inequalities in the criminal justice system, relying on databases and algorithms built on a history of discriminatory policing.
There is also evidence of facial recognition misidentifying black people, women and young people at higher rates than older white men, Cyril noted, arguing that the use of the technology against immigrants was unjustified: “We should not be criminalizing people trying to escape the travesties … in their countries of origin.”
Brian Brackeen, the CEO of Kairos, a face recognition company, noted that problems also arise when cameras take pictures through glass, and that “false positives” were more likely to affect people with darker skin: “It’s inevitably going to lead to problems.”
In records obtained by the Verge, officials said the cameras were capturing at least 1,400 vehicles over three days as part of the test programs. The lab deleted the images after the analysis, according to Gabris.
Asked how the government would store and use the images in the new pilot, Gabris said officials were still working on a “privacy impact assessment”, adding that the Department of Homeland Security has been testing biometrics since 2003. In 2016, Congress provided up to $1bn to CBP for this work. The agency “has the authority to capture scene images from all vehicles”, she added in an email.
Ebadolahi said she was concerned about how the government could expand this targeting, given that CBP has a wide geographical reach and considers the “border” to be anywhere within 100 miles of the country’s boundary lines: “They’ve been thwarted so far in building a physical wall. Now, they are trying to build a virtual wall. But the difference is a virtual wall can exist anywhere.”
If you have stories of government surveillance or targeting at the border, contact sam.levin@theguardian.com
