Technologists, travel executives, civil rights activists, and ethics experts met Tuesday in Seattle to debate expanding facial recognition technology at Sea-Tac Airport as part of the federal government’s broader push to screen travelers using biometric data.
The Port of Seattle hosted the panel discussion to learn about the advantages and concerns that have been raised over the controversial technology. The Port plans to implement new guidelines to govern the use of facial recognition software at Sea-Tac, though some biometric screening already occurs at the airport.
The Port convened experts in response to an aggressive push by U.S. Customs and Border Protection (CBP) to implement biometric screening at ports of entry across the country. CBP’s goal is to deploy facial recognition tech on all international commercial flights within the next four years.
The plan alarms civil rights groups, like the ACLU, who claim facial recognition technology misidentifies minorities and women and higher rates. Digital rights advocates are also raising questions about privacy and cybersecurity.
“What’s happening at Customs and Border Protection and the Transportation Security Administration continues to raise significant flags for immigrants and refugee communities and other communities of color,” said Richard Stolz, director of the immigrant advocacy group OneAmerica, during the panel.
The ACLU’s Shankar Narayan echoed his concern during the discussion.
“This such a powerful surveillance technology,” he said. “It super-charges the government’s ability to harm communities that are already disproportionately impacted by surveillance … The worst impacts of tech, in general, and surveillance technology, in particular, have always been felt by certain vulnerable and marginalized communities.”
Despite those concerns, CBP is charging full-steam ahead with its plans. Facial recognition technology is already used in at least 19 airports across the country, according to the Port of Seattle. CBP expects facial recognition software will be used to screen 97 percent of flights from the U.S. by 2021.
CBP began developing facial recognition technology to comply with a series of federal mandates enacted in the wake of 9/11. The laws instruct CBP to use biometric technology to screen international travelers entering and exiting the U.S.
The technology compares photos of passengers taken at airports with photos the government already has from various identification services. When passengers check-in for their flights, CBP pulls their government ID photos and puts them into a pool with the rest of the travelers on that plane. As passengers board, their photo is taken and compared with the pool. The technology is used to verify travelers are who they say they are and check their info against government watchlists.
“If there are 300 people on the flight, we find every photograph we have of those 300 people. Generally, that means we will have about 1,500 pictures because we have multiple photos of each passenger,” said CBP’s John Wagner, in a 2017 post on the department’s blog.
The photos taken of U.S. citizens at airports are discarded within 24 hours, according to Stephanie Gupta of the American Association of Airport Executives, one of the panelists at the Port event. She said that photos of non-U.S. citizens are discarded within 15 days unless those individuals are selected for additional questioning. Facial recognition technology has identified 185 people attempting to enter the U.S. using false credentials this year, Gupta said.
CBP first partnered with Delta Airlines to pilot the facial recognition program at the Atlanta International Airport in 2016. A year later, JetBlue became the first airline to board passengers using facial recognition software instead of boarding passes. The technology is only used on international flights. CBP said in 2017 that officials were identifying between 2-10 undocumented travelers per day using biometric screening.
Speaking at the Seattle event Tuesday, Delta’s director for passenger facilitation, Jason Hausner, said that the program has been a success. Fewer than 2 percent of passengers opt-out of biometric screening and it has increased boarding speed by about 10 percent during peak times, he said. Delta wants to deploy facial recognition at Sea-Tac as early as the fourth quarter of this year.
Passengers can opt-out of facial recognition screening at airports but the digital rights watchdog Electronic Frontier Foundation found it can be tricky to do so.
Facial recognition has become a lightning rod, particularly when it is used by law enforcement agencies. The ACLU and MIT conducted studies that show the software has the potential to amplify human biases. Amazon, a leader in facial recognition, says those findings could not be replicated when with the settings recommended for law enforcement applications in place.
Microsoft — another developer of facial recognition software — has called on the federal government to enact guardrails for the technology. As it stands, facial recognition is largely unregulated with the exception of a few local jurisdictions.
Microsoft sent its ethics strategist, Jacquelyn Krones, to the Port event Tuesday. She recommended that the Port implement “reasonable limitations to data sharing agreements” to prevent the data collected by facial recognition systems from being used in ways travelers do not expect.
“There’s a concern that once you are in a facial recognition system, it could be used anywhere,” Krones said. “That really depends on how data is shared and how responsible data controllers and processors are about those agreements.”
Several of the Port commissioners seemed skeptical of the new technology, including Courtney Gregoire, who also works as assistant general counsel for Microsoft. She pressed Delta’s Hausner on the airline’s agreements to share data collected through its facial recognition program.
“Every time you sign on to process data and give it to a third party, whether it’s with CBP or a vendor, you can impose limitations as to when they delete the data, where else the data goes,” she said. “Just for the record, everyone knows what happens when you don’t have a data-sharing agreement that has limitations. It’s called Cambridge Analytica.”
But airlines and federal officials who are bullish about the technology say it is optional, expeditious, and accurate. They also aren’t the only ones deploying biometric screening in airports. Sea-Tac authorized Clear to deploy fingerprint and iris-scanning kiosks throughout the airport back in 2016. Travelers pay an annual fee and surrender some biometric data to skip the security line. Today, Clear operates in airports across the country.
The Port of Seattle plans to review the information presented Tuesday and hold another meeting in October where commissioners will discuss new guidelines for the technology.