Citing concern over potential discrimination and dystopian misuse, the Somerville City Council in Massachusetts today voted to ban facial recognition software use by local police and city departments. This makes Somerville, which is located near Boston and has a population of 81,000, the second U.S. city to outlaw the technology. The ban follows the introduction of a facial recognition ban passed in San Francisco in May. That ordinance was written in part by the ACLU, which also helped craft the Somerville ordinance.
In addition to a ban on active use of AI by city departments, the ordinance outlaws use of data or evidence produced by facial recognition software systems in criminal investigations or legal proceedings. The local law does not have the scope to restrict facial recognition use by state or federal law enforcement.
However, the law refers to facial recognition as the “functional equivalent of requiring every person to carry and display a personal photo identification card at all times” and cites concern over false positive facial recognition matches for women, young people, immigrants, and people of color.
The ordinance was approved in an 11-0 vote — hardly surprising, as 9 of the 11 Somerville City councillors sponsored the legislation.
As facial recognition software begins to trickle into the workplace and consumer devices like the iPhone use facial recognition to verify purchases, police are also adopting the largely unregulated technology. With the ball already in play, governments are just now beginning to consider how facial recognition software should be used in society.
Local residents shared 98 written comments, plus letters of support from the ACLU and a letter in opposition from trade organization Security Industry Association, Somerville city clerk John Long told VentureBeat in an email. All letters received by the city support a ban.
“Any new security technology must be proven effective before taxpayer money is spent implementing it,” Somerville resident Dmitry Erastov said in a letter.
Founding editor of Logic magazine and Harvard University fellow Moira Weigel is a technologist married to a technologist and a Somerville resident.
“We know that such technologies enforce racism and cis-heterosexism and hand over huge amounts of sensitive information to enrich democratically unaccountable private entities. It’s no accident that San Francisco, another city with a high population of technologists who understand how these systems work, [has] also voted to ban them,” Weigel said.
Also this week: Police body camera maker Axon today pledged to keep facial recognition software from its devices. On Tuesday, the Oakland Privacy Advisory Committee in California endorsed the wording of facial recognition ban legislation being considered by the city council.
Brian Hofer is chair of the Oakland Privacy Advisory Commission and coauthor of the San Francisco legislation, but passage of a similar ban in Oakland encountered some initial delays, he told VentureBeat last month. The California State Legislature is also considering a facial recognition ban on police body cam footage.
A number of AI experts or privacy advocates who advise state and national lawmakers about the risks and opportunities associated with facial recognition software favor a moratorium or outright ban on the technology.
The Boston area plays a fairly prominent role in facial recognition policy debates. And the Massachusetts State Legislature is currently considering the Face Surveillance Moratorium Act, a bill sponsored by a Boston area lawmaker.
Representative Ayanna Pressley (D-MA), who represents Somerville and Boston in Congress, has in recent weeks participated in multiple House Oversight and Reform Committee hearings that saw bipartisan support for temporary limitations on facial recognition software use by law enforcement.
MIT Media Lab’s Joy Boulamwini, who testified before a U.S. House of Representatives subcommittee this week, led assessments that raised serious issues with leading facial recognition systems from companies like Microsoft, Face++ and Amazon’s Rekognition.