San Diego Police Department
A San Diego police officer / Photo by Adriana Heldiz

California’s ban on police use of facial recognition is set to expire on Jan. 1, meaning cops in the field could re-deploy the technology if their departments choose to next year.  

The ban went into effect in 2020 with support from criminal justice reformers, who warned the technology was racially biased and ripe for abuse, but the Legislature failed to take up the cause a second time.  

A bill in the senate that would have extended the ban came up two votes short this summer. As initially written, SB 1038 would have banned police camera use of facial recognition indefinitely, but at the last-minute, Sen. Steven Bradford of Los Angeles County, the bill’s author, agreed to another three-year moratorium.  

“That was still too much time for members to get on board,” said Christopher Morales, Bradford’s legislative director.  

San Diego’s delegation was split on the extension — Sen. Toni Atkins was a yes, Sen. Pat Bates was a no — and two of its other members sat out the vote.  

Sen. Brian Jones supported the initial ban starting in 2020 but abstained from SB 1038 “because the studies are still pending,” wrote Nina Krishel, his communications director, in an email. Sen. Ben Hueso’s staff said he was unavailable for comment and offered no explanation for why he declined to weigh in. Like Jones, Hueso supported the initial ban.  

None of the local police agencies surveyed by Voice of San Diego in recent days said they had immediate plans to bring the technology back in 2023. They also didn’t rule out the possibility of reviving the program in the future.  

Doing so in at least one jurisdiction — the city of San Diego — will be harder, though not impossible, thanks to new rules governing surveillance technology.  

Shifting Attitudes  

About two dozen state and local governments, according to Reuters, passed laws restricting facial recognition between 2019 and 2021, with momentum coming from the Black Lives Matter protests. Elected officials were also sustained by reports of Black men being falsely accused of crimes and other research undermining the technology’s reliability.  

The ACLU, for instance, ran the faces of California lawmakers through facial recognition software and found that one out of five were erroneously matched to another person who had been arrested. Sens. Bradford, Jones and Hueso were all mistakenly identified. At the time, Amazon, Microsoft and IBM had placed restrictions on the sale of their facial recognition tools to police, and Axon, a body-worn camera maker, said it would reject the use of the software on its products too.  

But as Reuters also reported, many of those same governments have since backed off their restrictions on facial recognition, citing a rise in crime during the pandemic. 

Elected officials have also pointed to research by the federal government claiming the accuracy of the technology has improved — although the ACLU disputes that racial bias has been eliminated. In public hearings, Bradford petitioned his colleagues to err on the side of caution. He and others warned that linking facial recognition in the field to disparate forms of technology and databases could have serious consequences in a democracy.  

“We should be very concerned about letting this kind of very invasive technology creep into our personal lives because once it gets in, it is very hard for it to get out,” said Sen. Sydney Kamlager, also from Los Angeles. “And I know that we’re talking about this as it relates to law enforcement, but I haven’t heard of any technology out there that self corrects itself or self regulates.”  

When arguing against SB 1038, a lobbyist for the California State Sheriffs’ Association offered a window into how the technology could be used long term. “With the rise in active shooter events and brazen thefts in public spaces continuing,” said Usha Mutschler, “bans on tools that could assist officers and protect public safety we believe is shortsighted.” 

Before the California ban went into effect, the Union-Tribune reported that the San Diego Police Department was the biggest user of the technology, scanning 14,050 faces between 2017 and 2018. The images were run through a regionwide database known as the Tactical Identification System, or TACIDS, which contained around 1.8 million mugshots for comparison.  

Dozens of agencies, not just local police, had access to the database, including Border Patrol. At its peak, the regionwide program consisted of 1,300 smartphones and tablets, responsible for more than 65,000 scans between 2016 and 2018.  

The program was rolled out more than a decade ago with funding from the U.S. Department of Justice and within a few years had attracted national media attention and congressional interest. Police officials pushed back against the initial ban, arguing that the technology was used for legitimate reasons — to identify people who’d been suspected of a crime.  

The New York Times reported in 2015 that police officers throughout San Diego County found a match in criminal records about 25 percent of the time, but the usage was not proportional. A researcher at the Georgetown University’s Center on Privacy and Technology testified in Congress in 2019 that San Diego police used face recognition “up to two and a half times more on African Americans than on anyone else.”  

Four years prior, consultants working for the Automated Regional Justice Information System, which oversaw the data-sharing, noted that 15 percent of women were targeted for “voyeuristic reasons” and 65 percent of teenagers were targeted for no reason at all.  

Although police lobbyists told California lawmakers that facial recognition was justified thanks to the countless crimes that had been thwarted, Bradford’s Senate Public Safety Committee staff said they hadn’t received any evidence to prove that the ban had “significantly hampered law enforcement efforts in the two years since it became operative.”  

If nothing else, losing the ability to use facial recognition in the field added time to a detention. As SDPD Capt. Jeff Jordon put it, “We have to still identify people.” Instead, he said, officers need to drive someone back to the station to figure out who they are, often with the aid of finger-print scanners.  

Still, none of the police agencies surveyed by Voice of San Diego said they tracked the use of facial recognition scans that led to arrests or prosecutions. Most of the examples we know about come from news articles or law enforcement testimonials made public through public records requests.  

Those documents show that, at least early on, immigration authorities were thrilled about the possibility of unmasking individuals who are undocumented. “This is the most awesome thing EVER!!!” reported one agent. “We went over to the holding cells this morning to try out the TACID software. Pulled out three aliens who were recent jail releases.” 

New Oversight in San Diego 

Bringing the technology back will be more difficult in the city of San Diego today than it was a decade ago because of the work of the community groups that make up the Trust SD Coalition and the measures taken up this year by the City Council.  

The city’s surveillance ordinance lays out the process. Officials need to come up with a use policy and impact report. They need to go out to various City Council districts where they intend to deploy the technology and submit their documents to the privacy advisory board. It then goes to elected officials for a discussion and vote. 

The ordinance highlights facial recognition as one particularly problematic tool necessitating greater oversight.  

“It’s this exact technology and others like it that inspired the need for the ordinance and communities to stand up for themselves,” said Seth Hall, a member of the Trust SD Coalition. “And as a result of passing it, San Diegans will have, at the very least, vigorous public debate about the use or acquisition of the technology by any city department.” 

There’s a caveat. The city’s ordinance, thanks to a 5-4 vote in June, exempts SDPD officers who work on federal task forces from disclosing the technology they use. Congress has not passed laws governing facial recognition, which federal agencies often rely on for a variety of reasons.  

“There are government agencies running around our town not subjected to the surveillance ordinance,” Hall said, “so San Diegans need to be aware that harmful technology like this could still be pointed at them.”  

Leave a comment

Your email address will not be published.

This site uses Akismet to reduce spam. Learn how your comment data is processed.