The FBI has a unit for facial recognition services that oversees a database with more than six hundred million photos. While the Government Accountability Office (GAO) recommended that the FBI take steps to ensure the accuracy of how the agency uses facial recognition, the FBI has failed to implement such changes.
Dr. Gretta Goodwin, the director of the homeland security and justice section of GAO, testified during a House Oversight and Reform Committee hearing. She told representatives the FBI collects photos from 21 state partners and two federal partners outside of the agency. However, the “FBI cannot know how accurate those systems are” because they still do not conduct adequate tests.
The hearing was the second in a series of hearings by the House Oversight and Reform Committee chaired by Representative Elijah Cummings.
In January, it was reported that the FBI piloted Amazon’s facial recognition software known as Rekognition. Officials confirmed “the pilot kicked off in early 2018 following a string of high-profile counterterrorism investigations that tested the limits of the FBI’s technological capabilities.”
Not long after, a coalition of civil liberties, human rights, faith, and racial and immigrant justice groups wrote letters not only to Amazon but also Microsoft and Google. They demanded the companies not sell “face surveillance technology to the government.”
MIT researchers revealed that same month that “Rekognition’s facial analysis feature mistakenly identified pictures of women as men and darker-skinned women as men 19 percent and 31 percent of the time, respectively. By comparison, Microsoft’s [service] misclassified darker-skinned women as men 1.5 percent of the time.”
Kim Del Greco, a deputy assistant director for the FBI’s Criminal Justice Information Services, was asked if the FBI was using or still piloting Amazon’s Rekognition software. She said the agency is not using the service. “We do not perform real-time surveillance,” Del Greco added.
The committee requested documentation from the FBI that showed Amazon’s service was not currently in use. Del Greco pledged to provide the information before the next hearing is held.
Massachusetts Democratic Representative Ayanna Pressley said, “It has been abundantly clear that facial recognition technology is flawed by design, unlawfully producing matches due to algorithmic bias, including to everyday Americans and, in fact, even members of Congress.”
In July 2018, the ACLU tested Amazon’s Rekognition software and found “the software incorrectly matched 28 members of Congress, identifying them as other people who have been arrested for a crime.”
One of the members mismatched was Representative Jimmy Gomez, a member of the committee. He asked Del Greco about biases in the technology, and she said FBI personnel are not trained on the potential inaccuracies of facial recognition algorithms.
Pressley applauded the city of Somerville, which is in her congressional district, because they passed an ordinance in May to ban the use of facial recognition technology until regulations for the technology are passed.
“I don’t think many Somerville residents would support a law requiring every resident to wear a visible identification badge on their chest every time they leave the house, but that’s exactly what facial recognition technology is,” declared City Council member Ben Ewen-Campen. “What’s more, given what we know about the troubling racial biases in facial recognition systems, and their chilling effect on free speech and community activism, I think it’s clear that Somerville needs to lead on this important issue by banning the use of facial recognition by our local government.”
During her questioning, Pressley addressed the key issue of how the Transportation Security Administration (TSA) has incorporated facial recognition technology. Passengers in airports are able to opt-out of a pilot program currently being conducted, but they may not be fully aware that they have to or can do so if they do not want to be screened with the technology.
The Electronic Frontier Foundation (EFF) suggests the reason why there may be confusion over opting out is the TSA has at least three checkpoints with facial recognition technology.
“Airlines want to use your face as your boarding pass, saying ‘it’s about convenience.’ CBP [Customs and Border Protection], which is part of the Department of Homeland Security (DHS), wants to use your face to check against DHS and State Department databases when you’re entering or exiting the country; and the TSA wants to compare your face against your photo identification throughout the airport,” according to EFF.
The digital privacy and civil liberties organization adds, “If people are upset now, they will be furious to know this is just the beginning of the “biometric pathway” program: CBP and TSA want to use face recognition and other biometric data to track everyone from check-in, through security, into airport lounges, and onto flights (PDF). They’re moving fast, too, despite (or perhaps because of) the fact that there are no regulations on this sort of technology: DHS is hoping to use facial recognition on 97 percent of departing air passengers within the next four years and 100 percent of all international passengers in the top 20 U.S. airports by 2021.”
Austin Gould, an assistant TSA administrator who oversees the deployment of facial recognition, was confident the agency is doing all it should to protect privacy, which is unlikely to calm anyone troubled by the growing use of the technology.
“This stuff freaks me out. I’m a little freaked out by facial recognition, Mr. Chairman,” Michigan Democratic Representative Rashida Tlaib declared.
Tlaib highlighted Project Green Light in Detroit, a program deployed in 2016 that relies upon real-time video surveillance. She said residents in the 13th congressional district have been “subjected to increased surveillance and over policing for decades.”
It started as a program for monitoring gas stations, liquor stores, and other businesses during the late evening. “The system has expanded to over 500 locations, including parks, churches, schools, women’s clinics, addiction treatment centers, and now public housing buildings,” Tlaib added.
The Detroit Police Department incorporated facial recognition into Project Green Light, which gives the police the ability to locate anyone with a Michigan driver’s license or an arrest record in real-time and cross-check it with a database of over 50 million photographs. The system likely violates the right to privacy by allowing police to conduct surveillance of citizens without a warrant approved by a judge.
“Has the FBI ever used facial recognition deployed at or near a protest, political rally, school, hospital, courthouse, or any other sensitive location?” asked New York Democratic Representative Alexandria Ocasio-Cortez.
Del Greco told Ocasio-Cortez the FBI had not. But facial recognition technology has been deployed at political protests.
“Marketing materials from the social media monitoring company Geofeedia bragged that, during the protests surrounding the death of Freddie Gray while in police custody, the Baltimore Police Department ran social media photos against a face recognition database to identify protesters and arrest them,” according to EFF.
Many of the law enforcement agencies that the FBI partners with may be abusing this technology. And, if they are abusing the technology, the FBI does not seem to believe it has an obligation to make sure those abuses do not influence how they employ facial recognition services at the federal level.