Page 6 of 6
Merritt's biggest concern is that even if PredPol works in predicting where crimes will happen, translating that into actionable strategies is another problem altogether. "Talking to L.A. made me proceed with caution. ... In L.A. I heard that many officers were only patrolling the red boxes, not other areas," says Merritt. "People became too focused on the boxes, and they had to come up with a slogan, 'Think outside the box.'"
At least one city has ditched PredPol because it lacked the basic resources to make use of it. The farming city of Salinas signed a three-year contract with PredPol in July 2012 for $75,000. Despite a reported 50 percent increase in the accuracy of violent-crime predictions by PredPol over Salinas police's previous hotspot crime-mapping method, the Monterey County Herald reported on Aug. 1 that police were too busy responding to a high volume of calls for service to adequately patrol the boxes highlighted by PredPol as likely areas for gun violence. Salinas Police Chief Kelly McMillan told the Herald that Salinas was the first time PredPol's software had been used to predict violent crime instead of property offenses, and that the volume of calls for service meant that officers could only patrol the PredPol boxes for six minutes at a time. And because there has been no independent analysis of PredPol's software in Salinas, or anywhere, it's not clear that it was even working.
Salinas amended its contract with PredPol on Sept. 22 after paying $25,000 for one year of services, effectively ending the software's use there. However, Salinas will still allow PredPol to access its crime data and has agreed to still do joint public relations on behalf of the company.
Rodriguez of the LAPD's Foothill Division believes that one of the reasons why his command has had success with predictive policing is the resources they are able to devote to policing the locations forecast as likely areas for crime. "One of the problems I've seen with other cities is that they buy into this program and think that'll be all they have to do — they expect results for the cost they pay," says Rodriguez. "This is not the panacea. We're dealing with statistics, we're dealing with probability," he says. "It's a big wide net we cast out into the ocean, and there's going to be some seepage."
That is, of course, if PredPol's version of predictive policing isn't an illusion based on incomplete science and aggressive marketing. American law enforcement's growing fascination with data-reliant strategies for complex and intransigent crime problems, says Crockford of the Massachusetts ACLU, is a troubling trend.
"It seems like we're moving in the wrong direction in how we think about crime," says Crockford. "Instead of figuring out why people are robbing houses, we're para-miltitarizing our police, turning all of them into robocops who take directions from computers as to how they go about their day." Crockford believes that relying on for-profit companies to deliver effective crime-fighting solutions poses serious risks. "There's a danger in overlap of the private sector and public sector. Policing shouldn't be influenced by corporate interests that profit from Big Data and that have an obvious interest in promoting these new technologies."