Police use of facial recognition technology usually only makes the headlines when it’s connected to a high-profile crime. But it’s increasingly being deployed in more mundane circumstances: There have been reports of police departments employing facial recognition to spot criminals from gas station security footage, identify suspects who aren’t carrying photo ID , and assist in social media surveillance . To many advocates, that’s a worrisome trend.
Lawmakers are starting to scrutinize the technology as well. Municipal legislators in San Francisco unveiled an ordinance proposal on Tuesday that, if passed, would make the city the first in the country to completely ban government use of facial recognition systems. The bill would also establish an audit system for agencies acquiring any surveillance technologies.
“The propensity for facial recognition technology to endanger civil rights and civil liberties substantially outweighs its purported benefits, and the technology will exacerbate racial injustice and threaten our ability to live free of continuous government monitoring,” the proposed legislation reads.
Privacy rights organizations have generally supported this proposal, pointing out that facial recognition systems often produce false positives for women and people of color, and that the technology could have a chilling effect on a number of civil liberties. Advocates fear that facial recognition could eventually endow the government with the ability to keep tabs on a person’s relationships, religious activities, political associations, and other personal details. While it may seem excessive to jump straight to a complete ban, privacy experts say it’s preferable to having few regulations at all, which is the case now.
“There certainly can be a debate on using this tech in more limited ways with clear checks, but the public and lawmakers should be deciding what uses should be allowed, not the other way around,” says Jake Laperruque, who serves as senior counsel at the Constitution Project and does work with facial recognition and privacy.
The lack of laws regulating facial recognition has allowed police departments, in particular, to apply the technology with little to no oversight or limits. Now, with activists, academic studies, and media consistently drawing the public’s attention to the technology’s dangers and deficiencies, localities and states across the country are considering whether to restrict its use.
While it may seem excessive to jump straight to a complete ban, privacy experts say it’s preferable to having few regulations at all, which is the casenow.
Indeed, while San Francisco would be the first city to have a total facial recognition ban, bills under consideration in Massachusetts and Washington state also propose moratoriums on the technology. But unlike San Francisco, these bans would last only until regulations of its use are passed in their respective state legislatures. Massachusetts’s SD 671 reads, in part, “the broad application of face recognition in public spaces is the functional equivalent of requiring every person to carry and display a personal photo identification card at all times, which constitutes an unacceptable mass violation of privacy.” Washington’s SB 5528 reads, “Before the widespread adoption of facial recognition technology by government agencies occurs, there must be a public discussion of acceptable uses of this technology, its accuracy must be demonstrated, and disproportionate impacts on vulnerable communities eliminated.”
When asked why the ACLU of Massachusetts, which has backed SD 671, decided to push for a temporary moratorium, Executive Director Carol Rose noted that a full-fledged ban did not seem politically viable in the state legislature. “It’s a way to create space to have a conversation about the permanent checks and balances needed to prevent abuse of the technology,” she said. Unlike the proposed ban in San Francisco, the Massachusetts and Washington bills allow for the possibility that the technology may have important, but limited, utility—particularly in emergencies, like helping police search for abducted children or locate imminently dangerous individuals in crowded areas.
But some advocates, including Rose, hesitate to make exceptions for it even in dire situations. Reserving the technology for emergencies still necessitates building databases containing millions of faces, the very existence of which would allow for the possibility of misuse. “Facial recognition and other kinds of biometric surveillance are really just a high-tech form of a watch list,” says Rose. That position also shows up in the proposed ban in San Francisco.
It also may be difficult to craft legislation that sufficiently silos the technology’s use, as opportunities for its abuse will likely multiply the more advanced and established it becomes. “We only have to look at the most recent generations of surveillance technology to see how those systems have been turned from one ‘beneficial’ purpose to another, such as tracking immigrants or spying on Black Lives Matter protesters,” says Matt Cagle, a technology and civil liberties attorney for the ACLU of Northern California. There have been reports that law enforcement used facial recognition to match people protesting the 2015 death of Freddie Gray with their social media photos, for instance.
Experts like Laperruque nevertheless acknowledge that legislators and the public could come to the conclusion that the advantages of using facial recognition in life-saving circumstances would outweigh the risks it poses to civil liberties. In that case, there would need to be strict checks and balances for government use, such as warrant approval and regular audits. “Having these sorts of incredibly powerful searches that can catalog people’s activities and find them in giant crowds, this is an over-powerful, overbearing tool that could be detrimental to democracy itself if left unchecked,” says Laperruque.
Of course, we may be getting a little ahead of ourselves. None of these bills have actually become law yet. A vote on San Francisco’s ordinance will come sometime in the next few months, while it may be a year or two before the Massachusetts bill will have a hearing during its legislative session. The Washington bill has been referred to the state Senate’s technology committee.
Rose is optimistic that the Massachusetts bill will be passed, noting that about two-dozen legislators from both sides of the aisle have already expressed their support. However, she worries that tech lobbyists may work to impede its progress. Companies like Google and Amazon , leaders in the facial recognition field, have sizable lobbying presences in the state according to public records.
And at AI Now—a symposium held October 16 about the intersection of artificial intelligence, ethics, organizing and accountability, presented by an NYU research institute of the same name—panelists warned that facial-recognition technology has troubling implications for civil rights, especially amid current debates about who has access to public space.
If one or more of these proposals do get passed by their respective legislatures, though, they could set an example for how other localities can start addressing the rapid proliferation of the technology.