Following its mandate to counter acts of terror and respond to related threats, the Bay Area Urban Areas Securities Initiative (UASI), a regional organization funded by a Department of Homeland Security grant program, has poured hundreds of thousands of dollars into automatic license-plate readers (ALPR) for local police across Northern California over the past few years.
But newly released numbers from the Bay Area’s UASI suggest that the vast majority of drivers’ location data, tracked and retained through ALPR programs, have not been linked to any terrorist activity or violent criminal activity.
ALPRs combine high-speed cameras with analytic image software, collecting the plate numbers of cars passing by specialized recording points. Those plate numbers are then typically compared against “hot lists” of people who law-enforcement agencies have interest in.
Last month, the Bay Area’s UASI released ALPR data from the Central Marin Police Authority showing that only .02% of the nearly 4 million license plates tracked over October of 2015 through April of this year resulted in matches to any police “hot list” databases. The data indicate that zero “known or suspected terrorists” have been tracked using ALPRs, and that only a handful of other matches related to other hot-list criteria.
Similar surveillance-to-threat disparities have been found in ALPR programs across the country. According to a 2013 American Civil Liberties Union report, only 47 of every million license-plate scans captured in Maryland and funneled to the state’s DHS-affiliated fusion center were linked with serious criminal cases involving car thefts, wanted persons, violent gang or terrorist organizations, sex offenders, or Maryland’s warrant-flagging program. Digging back further, the ACLU also found that, over January 2012 through March 2012, the Rhinebeck Police Department in New York had just a 0.01% hit rate for 99,771 plates scanned. Similarly, over August 2011 through June 2012, the High Point Police Department in North Carolina had only a 0.08% hit rate for 70,289 plates scanned.
Despite these small percentages of hits, however, law-enforcement officials say that ALPR technology remains essential for criminal investigations.
“If people could let me know which cars the criminals were driving, I would only look at those vehicles,” says Mike Sena, the director of the Northern California Regional Intelligence Center (NCRIC), which works with several local police departments using ALPRs. “The majority of the stats you see in that [Bay Area UASI] report are those that are just that quick hit, that quick alert… . But to me, the greater piece of that is giving the investigator a starting point when he comes up with that murder or violent crime… .”
In other words, ALPRs can give law enforcement important leads to follow when attempting to identify where suspects were at a given time. Sena notes, however, that it is impossible to quantify how often automatic license-plate readers help in this process, since detectives do not necessarily document every time license-plate data is used to determine the location of a persons of interest.
Civil liberties advocates, however, contend that holding such data on the movements of millions of people for extended periods of time (up to a year in the case of NCRIC-housed data) opens the door to Fourth Amendment violations or abuse of the data—such as the surveilling activists or ex-partners.
While there there are few publicized cases of that type of abuse—and little overall oversight of law enforcement use of this data—a Minneapolis Star-Tribune reporter showed just how easily abuses could occur by submitting a public-records request for police ALPR data and successfully tracking then-Mayor R.T. Rybak’s city-owned car.
“I understand why law enforcement would like to have as much info on everyone as possible. But whether that’s consistent with the values of America is a different story,” says Dave Maass, an investigative researcher with the Electronic Frontier Foundation, which advocates for civil-liberties protections in the digital world. “I think it is problematic when a law-enforcement agency collects evidence on everyone with the assumption that everyone may eventually be a criminal,” he says.
The spread of license plate trackers nationwide
Nonetheless, the use of license-plate readers at fixed locations or on roving police cars appears to have increased dramatically nationwide over the past decade. A 2014 report from the research organization RAND found that only 19 percent of agencies surveyed in 2007 reported the use of license-plate readers; by 2012 that figure had risen to 71 percent.
The federal government has helped several local police departments obtain ALPRs though counterterrorism grant programs. The data from the recent Bay Area UASI report, for example, came from three Central Marin Police Authority ALPRs, which were procured through a $132,553 UASI grant.
The Central Marin Police Authority obtained funding for the equipment citing national security concerns. Its application for UASI funding said that the ALPRs would close the department’s “capability gap” by addressing “roving criminals and possible terrorists operating in the Bay Area.”
But of the Central Marin Police’s relatively few “hot list” hits, nearly 99 percent were related to lost or stolen car cases. Privacy advocates say such data suggest the real potential for ALPR surveillance may be low-level-crime tracking, which could encourage disproportionate policing in low-income communities of color due to a perception that they are areas of higher crime.
The disproportionate impact of license-plate readers
Public-records-request data compiled by the Electronic Frontier Foundation in 2015 suggests that ALPRs attached to Oakland police vehicles primarily captured license-plate images in low-income neighborhoods. (Blue lines indicate image captures.)
Police use of ALPRs in Oakland also seems to leave neighborhoods with large white populations virtually untouched:
In 2013, Piedmont, a wealthy, majority-white town inside Oakland, spent almost $600,000 to purchase 39 license-plate readers to cover most of its Oakland border points of entry. The town cited safety concerns in its purchase. However, as Mike Katz-Lacabe of the Center for Human Rights & Privacy found, only 0.031 percent of the town’s 8.78 million captured license-plate images from December 2013 through June 2014 resulted in hot-list matches.
“The only thing that we’re really looking for are those major criminal activities,” says Sena of the Northern California Regional Intelligence Center, which houses license-plate data for municipalities like Piedmont. “We don’t want to get into the business of being alerted every time a guy drives down the street. We also want to give [law enforcement] the ability to focus on the major threats to our communities.” Sena does say that, contrary to his recommendations, local agencies can set license-plate-reader systems to alert on “whatever they want.”
This ability to cast a wider net through ALPR programs has played out far more controversially in other parts of the country. As the Electronic Frontier Foundation found, several municipalities in Texas, for example, have sparked controversy for allowing police to team up with private-sector companies to work like mobile debt collectors.
In places like Guadalupe County, the City of Orange, and the City of Kyle, Vigilant, a for-profit technology firm, gives police free license-plate readers and creates its own “hot list” for police, using police records on individuals with outstanding court fines. In exchange, police with license-plate readers identify drivers with outstanding fines during their patrols, offering them a trip to jail or the option to pay the original court fee (plus a 25 percent markup, all of which goes to Vigilant).
A Buzzfeed investigation earlier this year found this sort of punitive usage of ALPRs in Port Arthur, Texas, has helped fuel the disproportionate incarceration of black residents, who made up 70 percent of traffic citation arrests, despite only being 40 percent of the overall population.
“There’s not a lot of transparency on how they put together these hot lists or who is on the hot lists,” says Maass.
While these developments may help shore up municipal revenue bases, they can also incentivize local police to engage in extractive practices in poor neighborhoods, at odds with today’s increasing calls for community policing.
Automatic license plate readers “can shift police priorities,” says Maass. “Maybe you’re pulling people over who are on your hot list, rather than responding to calls for service or doing community policing… . In many ways, it’s law enforcement using technology to distance itself from the community, rather than engaging in partnerships with the community.”