San Francisco did it in May. Somerville, Massachusetts, in June. And on Tuesday evening, Oakland, California, became the latest to ban city departments — including police — from using facial-recognition technology.
As facial-recognition software becomes increasingly popular everywhere from airports to concerts to stores, concerns are on the rise about how well it works and the repercussions of when it fails. The technology identifies people from live or recorded video or still photos, typically by comparing their facial features with those in a database of faces (such as mugshots).
Yet while the technology could help with tasks ranging from solving crime to checking student attendance at school, it comes with fundamental privacy issues. Artificial intelligence researchers and civil rights groups such as the American Civil Liberties Union, are also worried about accuracy and bias in facial-recognition systems. They’re concerned that they are not as effective at correctly recognizing people of color and women; one reason for this may be that images used to train the software may be disproportionately male and white. No federal guidelines exist to limit or standardize its use, and few state rules are in place, leaving municipalities to decide for themselves what, if anything, to do.
Taking all this into account, some cities are deciding the risks of the technology far outweigh any possible benefits, and they’re working to outlaw it on their own rather than waiting to see what happens if and when it is deployed in their communities. Local governments are able to act nimbly and pass these rules relatively quickly.
They’re not trying to legislate how individual citizens use facial-recognition technology. In San Francisco, for instance, there is no prohibition against a person using a doorbell or security camera that can recognize specific people, and police could use footage from, say, a person’s Nest camera to assist in a criminal case. Yet these cities are drawing a line in the sand when it comes to government use of the software, and expect other cities to follow suit.
“I do think we’ll see more cities doing this. I think that’s very likely,” Rebecca Kaplan, Oakland’s city council president and sponsor of the ordinance to outlaw it, told CNN Business in advance of her city’s vote.
A unanimous decision to outlaw facial-recognition technology
Oakland’s city council passed the ordinance unanimously on Tuesday, and a second vote to finalize it is planned for September. Across the country in Somerville, just northwest of Boston, a similar measure also passed unanimously in late June.
As in San Francisco and Oakland, Somerville’s ban forbids city departments from buying or using facial-recognition technology for any purpose. Ben Ewen-Campen, the Somerville city councilor who sponsored the measure, said there was “overwhelming support” for the measure in the community, where there’s widespread concern about facial-recognition systems in particular and surveillance technology in general.
“We’ve gotten to the point [in the US] where these technologies are rolled out around the country with no transparency, no guidelines, regulations, and very little public understanding of actually how the technologies work,” Ewen-Campen said.
Misidentification is “not just a minor inconvenience”
In Oakland, Kaplan is similarly troubled by facial-recognition technology as it is today — particularly in the wake of studies that indicate a propensity to misidentify women and people of color. Oakland’s population is diverse, and Kaplan said the city has worked for years to remedy racial disparities in policing and the mistrust that is created in the community when police practices unfairly target people of color. Facial-recognition technology could make things even worse, she believes by misidentifying people as criminals.
“This is not just a minor inconvenience, if somebody is misidentified,” she said. “People have ended up shot and killed when they’re misidentified as a wanted suspect when they’re not.”
Kaplan had been “cautiously optimistic” Oakland’s city council would approve the ban; it was approved by the council’s public safety committee in June.
Just north of Oakland, Berkeley is contemplating its own prohibition of facial-recognition systems by city government. Kate Harrison, a Berkeley city council member sponsoring the ordinance, is also worried about misidentifying people, as well as the civil-liberties violations she believes the technology presents.
“We take people’s personal rights very seriously here,” she said.
Berkeley city council’s public safety committee heard testimony about the proposed rule on Wednesday; if it passes there in a future meeting, it will go to a city council vote.
Cities as technology regulators?
Some efforts by states and the federal government are underway to regulate the use of biometric technologies — including facial recognition and other body-based identification methods like fingerprint recognition — though few rules have passed (in Illinois, for instance, a law forces companies to get consent from customers before collecting biometric information).
A Senate bill introduced in March would force companies who want to use facial-recognition technology on consumers to first get their consent. At the state level, a Michigan bill under consideration would place a five-year moratorium on police using facial-recognition technology, while Massachusetts lawmakers are considering a moratorium on government use of any biometric surveillance system in the state. California, meanwhile, is considering a bill that would forbid police from using biometric technology like facial recognition in body cameras.
Yet cities and other local governments may be uniquely poised to challenge facial-recognition technology: they can quickly gauge the feelings of constituents — which include far fewer people than need to be considered for state or federal laws — and are often able to act quite rapidly.
In Somerville, for instance, Ewen-Campen proposed the facial-recognition ban in the spring as an addendum to a larger surveillance oversight draft policy that required approval from the city council and public discussion before deployment of surveillance technology. The ban was approved June 27.
“It’s certainly not easy to pass legislation in any context, in any legislature,” said Matt Cagle, a technology and civil liberties attorney at the ACLU of Northern California. “But local governments are naturally more responsive to their residents and they’re able to act quicker in many ways.”
Cagle expects pending facial-recogition bans in Oakland and Berkeley to pass, and other cities to consider similar rules in the near future. Already, according to an ACLU count, a dozen US cities have passed more generalized anti-surveillance laws in the past several years, including Somerville, San Francisco, Oakland, and Berkeley. A number of cities are considering their own such legislation, along with the states of Maine and California.
Additionally, city-level legislation is direct, says Evan Greer, deputy director of digital rights nonprofit group Fight for the Future. City councils and mayors have control over police department spending, so this kind of legislation is a way of saying no, she said.
“I think there’s something visceral about it for people that’s making it percolate up as something that needs to be acted on,” she said.
Not everyone’s a fan of the ban
Opponents of these bans, which include some members of law enforcement, are concerned that their passage could take away tools that could help protect public safety.
Oakland’s chief of police, Anne Kirkpatrick, said in a report to the city council that while the police department doesn’t use real-time facial recognition or plan to purchase the technology, it doesn’t think a ban is needed since the city already has a policy limiting the purchase of surveillance technology. She added that facial-recognition technology can let police speed up the “time-consuming manual process of connecting images from crime scenes to local mug shot databases.”
Instead of banning the technology, the department suggests banning only real-time facial recognition technology to send a message about concerns surrounding the technology, while leaving the door open for its eventual use.
Cory Salzillo, legislative director for the California State Sheriff’s Association, is also concerned about localized bans of the technology. City-specific rules could make it confusing for members of his group, which include the state’s 58 elected sheriffs, to do their jobs since city police departments may contract with sheriffs for traffic or patrol services, he said.
Like many proponents, he expects more cities and counties will consider whether to outlaw facial-recognition technology.
Ewen-Campen, who pushed Somerville’s ban, thinks this kind of deliberation is necessary.
“I think until state and federal government steps up, that is, in my opinion, what cities have to do,” he said.