'Predictive policing' isn't in science fiction, it's in Sacramento
Search for solutions
Surveillance and other technologies give police new tools to fight crime. But privacy advocates and civil liberties groups ask: At what price?
SACRAMENTO, Calif. — Officer Matt McPhail happened to be at his desk when the first alert went off.
A Nissan sedan had crossed the intersection of San Juan and Truxel where the Sacramento police had just placed one of two custom-built surveillance cameras. The system ID'd the vehicle as stolen.
“I said, ‘Hey, if anybody’s in the area, you know, keep an eye out for this car,’ ” recalls Mr. McPhail, a public information officer for the department. “And a helicopter was in the area and some officers went by and found it.”
That was 2014. The city has since installed 32 police observation devices, or PODs. Now Sacramento – like New York, Houston, Miami, St. Louis, and other cities before it – is looking at the next step: the launch in October of a “real-time crime center,” a central location from which officers could monitor all their existing surveillance technologies, PODs included.
The idea is that consolidating information about criminal activity – from stalking complaints to potential lone wolf terrorist attacks – would make law enforcement more effective at investigating and perhaps preventing some incidents. The process would also promote accountability and transparency at a time of rising tension between police and the black community, providing evidence of both police and suspect behavior during tense encounters, proponents say.
But the technology raises big privacy issues. Already concerned about PODs, privacy advocates are troubled by the prospect of centralizing law-enforcement data, especially in a post-9/11 world where data is being shared more widely across federal, state, and local lines. The technology is already causing a populist backlash.
“The theory of policing has changed,” says Rebecca Lonergan, a University of Southern California law professor who spent 16 years prosecuting public corruption and national security cases for the United States Attorney’s Office in Los Angeles. “There’s an understanding now that we really need to centralize all of our information.”
Protection for large gatherings
McPhail is usually off on Fridays, but he clocked in the day after Micah Xavier Johnson shot and killed five police officers in Dallas last month.
That same day, McPhail explained how the city’s new crime center would help prevent future tragedies in large public gatherings. The department is planning to install 10 new cameras at the city’s Golden 1 Center arena. The crime center itself will have banks of state-of-the-art screens showing live, high-definition video feeds of major traffic and transportation sites, as well as trained personnel who can provide real-time information to officers in the field and use data analysis tools to interpret any data being collected.
“It’s not a replacement for old-fashioned police work,” McPhail says. But “this is kind of like a natural progression … of how we do business, the questions we would ask. It’s like the first bread crumb along a trail where at least our investigators know where to start looking.”
It's called predictive policing, and law enforcement agencies in other cities are already making it happen. In New York City, home to the oldest and arguably most sophisticated real-time crime center in the country, police can use surveillance and data analysis technology to identify suspected criminals or terrorists based on anything from a birthmark to a limp.
The idea is that people who have a record get their identifying marks loaded into a database. So if police need to identify someone they see, they can type in the visible characteristics and then get that person's name and information.
With terrorism, “it’s so hard to find those few [radicals] that really are serious about it,” says Professor Lonergan at USC. “The only way you find them is by doing the kind of data collection and data mining that we’re talking about.”
The technology can also be used for more routine policing, such as addressing stalking claims. If complainants have a license plate number and are able to give at least three places where they might have seen the stalker, the POD system can be queried to see if the vehicle was there, McPhail says.
“Now I might have a stronger case to substantiate a stalking claim in advance of something potentially much more serious happening to our victim," he says. "And it has been used to that effect.”
More policing not the answer?
For some privacy and civil liberties advocates, such predictive strategies in routine police work is a problem, not a solution.
“There’s a shift in the primary modality of policing, where it’s not just the old investigating methods being employed, but preemptive policing based on hunches,” says Hamid Khan, a coordinator with the Stop LAPD Spying Coalition, an alliance of community groups that aims to prevent undue surveillance of marginalized communities in Los Angeles. “It’s become part of a larger architecture of surveillance.”
He and other critics say that sort of predictive policing reinforces racial profiling and violates civil liberties, with little accountability on the part of the officers who employ such methods. Worse, the strategy fails to address the underlying reasons for which people often commit crimes.
“[T]he deepest flaw in the logic of predictive policing is the assumption that … what the model predicts is the need for policing, as opposed to the need for any other less coercive social tools to deal with the trauma of economic distress, family dislocation, mental illness, environmental stress and racial discrimination that often masquerade as criminal behavior,” writes Aderson Francois, a professor of law at Howard University in Washington, in an op-ed for The New York Times.
Law enforcement should stay out of the surveillance and data collection business, critics say – at least, until lawmakers are able to develop clear policies that regulate their use.
“Technology can be liberating or it can be a tool for control,” notes Shahid Buttar, director of grassroots advocacy at the Electronic Frontier Foundation (EFF), a civil liberties nonprofit based in San Francisco. Conscientiously developing and implementing policy to govern that technology and its use, he says, could spell the difference between the two.
Take body cameras, Mr. Buttar says. More police departments are adopting the devices, hailing them as a tool for improving transparency and accountability, and rebuilding trust in communities. Yet a recent report from the Brennan Center suggests that some departments continue to struggle with privacy concerns months after making body cameras a requirement for officers.
In Dallas, for example, officers don't need to inform residents they are recording as long as they are "in a private residence in an official capacity." And in Baltimore, homes are treated the same as any other property, though officers "have discretion not to record in 'sensitive circumstances,' " according to the report.
Some drivers are staging their own protests, snapping up anticamera license-plate covers, even sprays, that keep their plates visible but difficult to photograph.
New county law
A couple of interstates down from Sacramento, Santa Clara County passed an ordinance that tries to address these concerns. The first of its kind at the county level, it requires the sheriff's department and the district attorney’s office to inform the Board of Supervisors and the public before they purchase any new technology. They also have to lay out any privacy and due process implications of the devices, develop a set of policies that would regulate their use, and submit an annual report on the technology’s use to the board.
“This notion that we have to choose between public safety and privacy protection is wholly false,” says Joe Simitian, a county supervisor and former California state senator. “We can protect the public and respect their privacy and due process rights at the same time.”
Privacy advocates have praised the ordinance.
“What protects privacy is a straightforward explanation of what we’ve got, what we’re using, when we’re going to use it, and what the boundaries are,” says Tracy Rosenberg, executive director of Media Alliance, a nonprofit media group that aims to defend press freedom and civil rights. The Santa Clara County ordinance does just that, she says, and should be raised as a model for other local legislatures to look into.
Limiting the cameras
Standing at the corner of Fruitridge Road and Franklin Boulevard on a sunny July day, McPhail points to a traffic light, where three small, domed cameras jut out from a beige metal box marked with the logo of the Sacramento Police Department. Sacramento does not have anything in the books resembling Santa Clara County’s new ordinance, but McPhail points out that all three cameras face only the streets, avoiding even the businesses that line the two main roads.
The department also works closely with community leaders through its “Cops and Clergy” program, McPhail says.
And while he understands concerns over data collection – footage from the PODs is stored in a third-party database and is accessible to Sacramento police for up to two years – he says that for his city, at least, residents have little to worry about.
“It's not as if there's somebody sitting here waiting for a particular [person] to walk by everyday and using [the cameras] as some means to follow people or monitor their daily lives. That's not what it's for,” he says. “We don't have time for that.”
“People do value their privacy,” he adds. “Even police officers value that. When I'm at home with my family, I don't want to have to watch what I'm doing or be cognizant of where I'm having to go out of concern that somebody … might be watching me.
“But,” he says, “I think there's also a recognition of the value that this can add, in terms of us having a positive impact on the lives of people in these areas with reduction of crime, ability to respond quickly to things that are occurring, and the potential to solve crimes that might otherwise be unsolvable.”