Is Predictive Policing the Future of Law Enforcement?

Los Angeles Police Chief Charlie Beck has seen the future of policing, and it lies in predicting crime before it happens.

'Predictive policing' is suddenly popping up everywhere, with computer scientists exploring models that predict crime and departments like Beck's LAPD jockeying for federal grants to explore the potential of predicting the future.

The L.A. Times explored the controversial issue in a long piece on Saturday. And while there's plenty to be concerned about with this trend, there's also a hope that it could lead to smarter law enforcement. Like many law enforcement tools and innovations -- it could be a huge benefit, if we use it right.

For at least 15 years, police have been mapping crime and crunching data using systems like CompStat, and this data-based policing quite likely played a role in the recent declines in violent crime. We know better where to deploy patrols and when, and therefore we prevent some crimes of opportunity. Predictive policing is the logical next step. Using computer models and data on previous crimes, physical environments and more, scientists say they'll be able to predict crimes, helping police anticipate criminal activity and prevent it.

I'm a big fan of using data to improve our criminal justice system, and law enforcement agencies are right to explore the possibilities here. But a few red flags pop up, too.

- The hopes for predictive policing are a bit unfounded. Will the computer models work, or will we spend millions of dollars and officer-hours on wild goose chases? It's uncertain whether the complex, unpredictable behavior of human beings can be predicted by a computer model.

- Will these predictions simply increase police presence in poor urban neighborhoods? Residents already deal with an almost constant police presence (and, sometimes, undue suspicion). To some extent, of course, police should increase patrols in high-crime neighborhoods, but there's a limit. If predictive policing simply increases the police presence in these neighborhoods, it may pull cops away from other important work. Which brings me to my next point:

- We shouldn't discount the value of a police officer who is engaged in his or her community and knows the lay of the land. Maybe the money going to predictive computer models should instead be focused on innovative violence-prevention programs like CeaseFire in Chicago or Aim4Peace in St. Louis.

Unlike a proposal in Florida to start "rehabilitating" kids based on crimes they haven't committed yet, this form of predictive policing doesn't tread too heavily on civil liberties. Essentially, we're talking about interpreting patterns of data to wisely deploy resources. It has some potential.

One of the most interesting exchanges in the L.A. Times article comes from UCLA anthropologist Jeff Brantingham, who is helping run the school's exploration of predictive policing in partnership with the LAPD.

"The naysayers want you to believe that humans are too complex and too random — that this sort of math can't be done," Brantingham told the L.A. Times. "But humans are not nearly as random as we think, In a sense, crime is just a physical process, and if you can explain how offenders move and how they mix with their victims, you can understand an incredible amount."

Hat tip: The Crime Report / Image Credit: mysouthborough