On Tuesday, San Francisco became the first city in the U.S. to ban police and other city agencies from using facial recognition technology, which has wrongly tagged innocent people as criminals. The Board of Supervisors voted 8 to 1.
“We want to put the facial recognition technology, we want to put that genie back in the bottle. There are many ways to make our society secure without living in a security state. We have very good policing, but we don’t want to live in a police state. And this legislation tries to strike that down,” San Francisco Supervisor Aaron Preskin said on Tuesday.
Dave Maass, a senior investigative researcher at the Electronic Frontier Foundation, tells Press Play that during one hearing in San Francisco, a city staff member said that the technology would be able to predict crimes with 99% accuracy. “If somebody tells you that any technology can predict crime with 99% accuracy, they are 100% wrong.”
This tech has been used in other places too, including Los Angeles, airports throughout the U.S., and China.
Maass emphasizes that the technology is being used differently now. “Traditionally we’ve looked at face recognition as something that is used to confirm people’s identity. Maybe somebody is arrested, and you want to make sure they’re the person who appears in a previous mugshot. But that’s not what we’re looking at over the next few years. We’re looking at a world in which real-time face recognition can be applied to traditional surveillance cameras or body-worn cameras in order to identify people going around in their daily lives, or even tracking people’s movements across a city.”
Maass says more and more of these cameras are internet-connected, and companies like IBM and Amazon are looking at the footage to mark people and their movements.
So while surveillance used to be passive (for example, a camera would record you coming in and out of a store), it’s now active. It records who you are and where you are, then builds a long-term database of your whereabouts that the government and law enforcement agencies could access.
Facial recognition in Los Angeles
Maass explains that as far back as 2008, the LA County Sheriff's Department acquired devices that could recognize faces, tattoos and fingerprints, and then matched that information with mugshots; and in 2005, the LAPD asked to apply this tech to CCTV cameras.
But the LAPD hasn’t always followed the rules. Maass says that with license plate readers, which are already tracking us, agencies are required to put their policies online for everyone to see. “But we're looking at three or four years now that LAPD hasn't even complied with those basic requirements.”
Facial recognition at the federal level
The FBI, TSA, Customs and Border Protection are all using facial recognition.
Despite actions by the federal government, Maass says that any decisions related to facial recognition should be made locally.
“It's important to note that there are different types of crimes that the FBI and DHS are investigating than what is happening at LAPD. Maybe the FBI is looking at kidnappings or bank robberies or terrorist attacks,” he says.
“But you have agencies like in Sacramento, the agency that does welfare benefits was using license plate readers to spy on people who may have accepted more food stamps than they probably should have… What we end up seeing is using this technology to look at very low level crimes at the end,” he says.
Marketing and closed-door decisions
Authorities used this tech to identify the suspect in the mass shooting at the Capital Gazette newsroom in Annapolis, Maryland last year. And the San Francisco Police Officers Association said this tool isn’t foolproof, but can be helpful if safeguards are in place.
Facial recognition is not a foolproof technology, but it can be a helpful tool if safeguards are in place.https://t.co/Q3yT0awxaX
— San Francisco POA (@SanFranciscoPOA) May 15, 2019
Maass counters, “These are often talking points that we see in marketing materials by the vendors of this technology. And oftentimes you will hear exact verbatim text in an advertisement that then comes out of the mouth of someone in a police union.” He’s concerned that many decisions around this tech are made in closed rooms between a salesperson and a police department.
“One of the good things about the San Francisco ordinance is in addition to the face recognition ban, it ensures that any acquisition of surveillance technology has to go through a public process, and elected representatives have to decide whether it's appropriate for the community or not,” he says.
--Written by Amy Ta, produced by Adriana Cargill