In November, voters in Bellingham, Washington, passed a ballot measure banning government use of face recognition technology. It added to a streak of such laws that started with San Francisco in 2019 and now number around two dozen.
The spread of such bans has inspired hope from campaigners and policy experts of a turn against an artificial intelligence technology that can lead to invasions of privacy or even wrongful arrest. Such feelings got a boost when Facebook unexpectedly announced on the day of the Bellingham vote that it would shutter its own face recognition system for identifying people in photos and videos, due to “growing societal concerns.”
Yet a few months earlier and about 100 miles from Bellingham, the commission that runs Seattle-Tacoma International Airport passed its own face recognition restrictions that leave airlines free to use the technology for functions like bag drop and check in, although it promised to provide some oversight and barred the technology’s use by port police. SeaTac is one of 200 US airports where US Customs and Border Protection uses face recognition to check traveler identities.
At least seven states adopted face recognition to verify the identity of people applying for assistance such as unemployment benefits. Even Facebook’s headline-grabbing shutdown of its face recognition features came with a caveat: The company said it will retain the underlying technology, because it might be useful in the future as a way to unlock devices or secure financial services.
This is the paradox of face recognition in 2021: The technology is banned in some places but increasingly normalized in others. That’s likely to continue, because face recognition is unregulated in most of the US, as there’s no federal law covering the technology.
Many uses of face recognition have lower stakes than in policing; some, like unlocking a phone with a glance, can be seductively convenient. Despite concerns about the consequences of errors and evidence that some systems perform less well on people of color, the technology has become easy for non-tech companies to access and is generally reliable if deployed with care. A 2019 report by the National Institute for Standards and Technology said the majority of commercial algorithms tested showed unequal performance on different demographics, but also that any differences were minimal or undetectable for some of the most accurate and widely used algorithms.
Apple’s Face ID phone unlock system may be the most widely deployed and used face recognition system, but US airports are forerunners in normalizing its use in public spaces and interactions with the government.
CBP first deployed the technology in 2016 in partnership with Delta Air Lines at Hartsfield-Jackson Airport in Atlanta to check the identity of people boarding international flights. The program has steadily expanded since, but it accelerated in 2021, in part because the agency determined that touchless technology was more valuable during a pandemic.
At the end of 2020, CBP had implemented face recognition gates for incoming travelers at 17 airports. This year it added the technology at 182 airports, which the agency estimates will cover 99 percent of inbound air travel to the US. The program stems from legislation passed in the wake of the 9/11 attacks requiring biometric identity checks for anyone entering or exiting the US. Facial recognition is used to check outgoing international travelers at 32 US airports. CBP says it has processed more than 100 million travelers using face recognition and prevented more than 1,000 “imposters” from entering the US at air and land borders.
Most PopularThe End of Airbnb in New YorkBusiness
Since 2018, Delta has worked with CBP to offer international passengers flying from Atlanta the option of checking in and going through security using face recognition instead of conventional documents. In 2019, the airline used face recognition during boarding for 86 percent of its international departures from Atlanta; the proportion fell during the pandemic due to modified boarding processes, but is now at more than 60 percent of international flights and rising. Delta recently expanded the program to allow domestic passengers with TSA Precheck departing from Atlanta to progress from check-in to boarding using only their face for identity. The airline built the new system in collaboration with the Transportation Security Administration, CBP, and travel security company Pangiam, and it plans to roll it out at other airports, starting with Detroit.
Ranjan Goswami, Delta’s senior vice president of customer experience, said the new process in Atlanta makes travel more convenient for passengers and is “a blueprint for the future.” The program is voluntary, and Delta does not save or store any biometric data, Goswami says.
Shaun Moore, a Pangiam executive who joined the company when it acquired his face recognition startup Trueface earlier this year, says the debate about police use of the technology can obscure its value in other areas. “It paints the industry a little unfairly,” he says. “While talk around regulation for law enforcement use shakes out, we’ve focused on areas where there’s less concern and less risk and people are getting comfortable.”
Moore says Pangiam offers its technology to federal law enforcement but not to state and local departments, and that he supports regulating law enforcement use of face recognition. The Air Force also uses Pangiam’s technology to speed identity checks at base entrances, and the cryptocurrency exchange Everest uses it sign up new customers.
Finance companies are also showing interest in face recognition to speed identity checks. Incode, an identity verification startup based in San Francisco, says its face recognition checked more than 140 million identities in 2021, roughly four times as many as in the previous three years combined. The company’s customers include HSBC and Citigroup, and it recently raised $220 million in funding from investors including JP Morgan.
Caitlin Seeley George, a campaign director at nonprofit Fight for the Future, finds the spread of face recognition in airports and other areas of daily life concerning. “We need to ban all facial recognition, because the harms of this technology far outweigh any benefits,” she says.
George considers seemingly benign or careful uses of the technology dangerous because they help normalize collection of personal and biometric data that can be hacked or exploited. “The more places people see it, the more comfortable people feel,” she says. “When we do things for convenience we may not be thinking through all the repercussions.”
At the same time, George is optimistic about containing face recognition. She points to Facebook’s decision to shut its tagging system, the spread of local bans, and legislation introduced to both houses of Congress this year by a group of Democratic lawmakers and Senator Bernie Sanders (I-Vermont) that would ban use of face recognition by federal agencies. Similar bills were introduced in 2020 but did not proceed to a vote.
Updated, 1-3-21, 8pm ET: An earlier version of this article incorrectly said Pangiam does not offer its technology to law enforcement.
More Great WIRED Stories📩 The latest on tech, science, and more: Get our newsletters!David Attenborough’s unending mission to save our planetFuture hurricanes might hit sooner and last longerThe Android 12 privacy settings you should update nowIt’s time to reimagine the future of cyberpunk25 amazing gift ideas under $25👁️ Explore AI like never before with our new database🎧 Things not sounding right? Check out our favorite wireless headphones, soundbars, and Bluetooth speakers