Can License Plate Readers Really Reduce Crime?

In March, police in an Atlanta suburb embarked on a surveillance experiment in the hopes of reducing crime in the area. Along public roads near the local Six Flags amusement park, officers from the Cobb County Police Department installed 13 solar-powered automatic license-plate readers from Flock Safety, an Atlanta-based startup on a mission to “eliminate nonviolent crime.” The company’s cameras, which typically cost around $2,000 a year each to rent and operate, were loaned to Cobb County free of charge.

Law enforcement officials chose the area, known as Zone 2, Beat 215, because it’s experienced disproportionately high amounts of property and violent crime, says Stuart VanHoozer, Cobb County’s deputy chief of police. During the first six months the license plate readers were installed, VanHoozer says he observed a dramatic change: The number of reported crimes like robbery and nonresidential burglary dropped over 50 percent apiece compared with the same period the year before. Between March and August, the department recorded 50 instances of “entering auto,” Georgia’s term for breaking and entering into a vehicle, compared with 138 over the same time in 2018. “It was not a decrease that we expected to see,” says VanHoozer.

On the surface, it appears as though a simple fix—installing relatively discrete license-plate readers—had an enormous positive impact. That’s the narrative Flock Safety has put forward. The company proudly touted the results of the Cobb County pilot in a press release it sent to WIRED this week, and advertises on its website that it solves “up to five crimes an hour.”

But experts say it’s not that simple, and that establishing a causal relationship between any given variable and fluctuating crime rates is no easy task. “I am not saying that the readers did not have an effect on crime—it is just that we cannot attribute any reduction in crime to the readers themselves,” says Alex Piquero, a professor of criminology at the University of Texas, Dallas.

Even police agree. “To make it very clear, we are not 100 percent positive that Flock cameras were the difference,” notes VanHoozer. “What we did see, though, is an incredible decrease in crime, starting when we put these cameras down there.”

Photograph: Flock Safety

Police departments have been using surveillance tools like automatic license-plate readers for years. They’re often supplied by technology vendors like Vigilant Solutions, which markets them directly to law enforcement. Flock Safety—a former Y Combinator startup backed by Peter Thiel’s Founders Fund—is one of several newer technology firms selling surveillance tech to private citizens. Josh Thomas, the head of marketing at Flock Safety, says the company generally targets groups like homeowners’ associations, whose members pay for the readers to be installed in their communities. Another company, Ring, which was acquired by Amazon last year, sells doorbell cameras for homes and businesses.

Both Flock Safety and Ring have worked with law enforcement, and the latter’s relationships with hundreds of police departments across the country, in particular, has drawn scrutiny from civil liberties groups and others. Advances in surveillance technology have spurred an ongoing debate over privacy and security, and whether individual trade-offs are justified in the name of public safety. Flock and Ring sell consumers on the idea that their products have the power to not only catch criminals, but also deter them from offending in the first place.

In a 2015 pilot program with the Los Angeles Police Department, Ring said the presence of its cameras reduced burglaries in neighborhoods by as much as 55 percent, a figure that has been disputed.

Is there something about license plate readers you think we should know? Contact the author at louise_matsakis@wired.com or via Signal at 347-966-3806. WIRED protects the confidentiality of its sources, but if you wish to conceal your identity, here are the instructions for using SecureDrop. You can also mail us materials at 520 Third Street, Suite 350, San Francisco, CA 94107.

“It’s in the interest of the jurisdiction and the manufacturer to advance the notion that this is the latest and most sophisticated technology,” says Elias Silverman, professor emeritus at John Jay School of Criminal Justice and the Graduate Center, City University of New York. “One needs to take these advances with a grain of salt and acknowledge they need to be studied by others.”

Flock’s cameras are now used in over 400 communities in 35 states, and have been credited with helping police solve a series of serious cases. But the overall effect installing license plate readers has on crime rates still isn’t clear, and likely can’t be determined by conducting a short experiment. Some studies indicate the readers don’t deter crime, while others have found the devices can potentially reduce certain types of offenses.

Maria Cuellar, a professor of criminology at the University of Pennsylvania who researches the use of statistics in the law, says pilots like the one in Cobb County only provide before and after comparisons, which alone can’t prove a causal relationship. “The problem with these is that so many things could have changed between the ‘before’ period and the ‘after’ period,” she adds. That includes everything from the number of cars passing through the area to broader demographic changes. The study was also relatively short. “With such a small sample size in terms of time, any changes could likely be noise rather than an actual signal,” says Cuellar.

Wider trends too have to be taken into account. Police say crime is down overall in Cobb County, as well as in nearby Atlanta. “We do believe that there are other things we are doing that have attributed to the general decline in crime,” says VanHoozer. He notes there are also social factors that might be contributing to the drop, like low unemployment.

Photograph: Flock Safety

Another possibility is that Flock’s license plate readers helped police apprehend a relatively small group of criminals who were responsible for the bulk of the crime in the area. In other words, rather than deterring people from committing an offense in the first place, the license plate readers may have helped apprehend repeat offenders.

Cops have used license plate readers for at least a decade, but the ones made by Flock Safety are arguably more powerful. They automatically catalog a vehicle’s model, color, make, and any distinguishing marks, as well as the date and time they passed through the neighborhood. The cameras ping law enforcement the minute a known stolen vehicle crosses their path, a feature VanHoozer says has been particularly useful in Cobb County. The Flock LPRs are even capable of detecting people walking by, and whether they have a dog in tow.

There are safeguards in place for how Cobb County police can use the cameras. Flock Safety only allows footage to be retained for 30 days, and according to department policy, officers need to document a legitimate law-enforcement reason and case number to access the tapes. But Flock Safety’s license plate readers are also used by private individuals, who aren’t subject to the same oversight. It’s possible a homeowners’ association may use the readers to improperly spy on or surveil its neighbors, for instance.

Even in the hands of police, license plate data has been used for controversial purposes, like by US Customs and Immigrations Enforcement to apprehend undocumented immigrants. And there’s always the risk the underlying law-enforcement databases the readers rely on may not be accurate. According to one estimate from the Northern California Regional Intelligence Center, the readers can be wrong around 10 percent of the time.

Last year, two men driving in the Bay Area said they were pulled over and handcuffed as police pointed guns at them, after an automatic license-plate reader incorrectly identified their rental car as stolen.

As license plate readers are installed in more communities, citizens will need to weigh the potential privacy risks against the promise they will reduce crime in their neighborhood. They should also keep in mind that a whole host of things can impact crime rates, and while surveillance technology is an increasingly common variable, it’s not the only one.

Is there something about license plate readers you think we should know? Contact the author at louise_matsakis@wired.com or via Signal at 347-966-3806.