The defenders and promulgators of data-driven, predictive policing — which is meant to anticipate crimes before they happen — face a PR problem: reassuring the public against fears that such methods are ushering in a totalitarian future reminiscent of the science-fiction film Minority Report.
Concerns about preemptive crime fighting through data hoarding and analysis are hard to assuage, however, because they are perfectly valid.
A lengthy feature published in the Guardian on Wednesday looked at the permeation of data-driven analysis in the LAPD and other municipal police forces. “As the ability to collect, store and analyze data becomes cheaper and easier, law enforcement agencies all over the world are adopting techniques that harness the potential of technology to provide more and better information,” it noted. “But while these new tools have been welcomed by law enforcement agencies, they’re raising concerns about privacy, surveillance and how much power should be given over to computer algorithms.”
The Guardian’s report describes an LAPD war room full of video screens. They show incidents of crime in real time; multiple newscasts; the seismic effects of earthquakes; and sections of the city as small as 500 square feet where algorithmic data-crunching indicates that crimes are most likely to take place.
At first glance, such systems seem benignly empirical. Why wait for a robbery or a shooting when algorithms working beyond the capabilities of human intuition can help prevent these incidents in advance? But such an understanding wrongly assumes the neutrality of information. The picture of crime to come is based on pre-existing police data, which we know to be biased and flawed.
Consider the background of this trend. During Bill “broken windows” Bratton’s first tenure as police commissioner in the mid-1990s, the NYPD introduced a system called CompStat. It wasn’t a computer system but a statistical management tool that processed crime and arrest reports and other police records. In many ways, it amounted to the birth of predictive policing.
The idea was that data wouldn't lie, and that a close look at this honest, pure data could prevent crimes before they needed to be solved. Of course, no police business — on the streets or in the stats — is so pure. Data theorist and artist Ingrid Burrington has called the false promise of neutrality an “ideology of data,” noting that the methodology behind information collection and analysis can distort the picture of reality they represent.
The NYPD’s targeting of poor and black neighborhoods for easy quota-filling outfitted CompStat with inherently biased data. CompStat’s numbers in turn justified the quota-chasing and underpinned a discriminatory stop-and-frisk practice that targeted statistical hot spots. The not-so-neutral data helped establish a self-fulfilling system of profiling that branded entire communities “criminal.”
It’s no coincidence that the most policed areas in American cities are the highest crime areas. A “crime” is, basically, what and where the police say it is. The resulting statistics only serve to reinforce the focus, which has a criminalizing effect on residents of the area by processing them within the criminal justice system from an early age. The result is a police practice that essentially keeps whole locales and minority groups in everyday opposition to law enforcement. If predictive policing were effective in ending crime, things would look differently.
“This is not Minority Report,” P Jeffrey Brantingham, a professor of anthropology at UCLA who helped develop the predictive policing system that is licensed to dozens of police departments under the name PredPol, remarked to the Guardian of his brainchild. “Minority Report is about predicting who will commit a crime before they commit it. This is about predicting where and when crime is most likely to occur, not who will commit it.”
Brantingham’s distinction between policing a place and policing a people speaks volumes. Perhaps his system doesn’t target particular individuals as likely perpetrators of crime. But places are not stopped, frisked, and detained — people are. And places are not vacant terrains where crimes do or do not happen. They are homes, neighborhoods, communities — they are populated.
Policing where and when a crime is most likely to occur entails predicting what sort of people might commit it. It’s akin to the profiling entailed in the Obama administration’s use of “disposition matrices” to target drone strikes. Consistently, specific individuals are not the targets of counter-terror drone strikes. Instead, problematically, areas and behaviors and demographic factors are used to determine where drones should strike.
As Brantingham noted, predictive policing hasn’t yet amounted to a _Minority Report_-style nightmare. But something equally pernicious is at play: a system of policing under the pretense of neutrality, which assures that certain oppressed groups remain designated and punished as criminal.
Follow Natasha Lennard on Twitter:__@natashalennard