What if police could forecast crime the same way meteorologists talk about the chances of scattered storms this afternoon?
Our Birmingham criminal defense attorneys know they are already trying.
While police work in the past has typically been about responding to a crime after it occurs, what law enforcement agencies are attempting is to respond to where crime is likely to happen before it happens.
You would be forgiven for thinking this sounds a bit like something out of a futuristic science fiction novel. In reality, it’s being done with a computer programming model called PredPol. It’s already being used in Los Angeles and Seattle, and is typically used to target certain types of crimes – say gun violence, for example.
You may be familiar with CompStat, a system developed in New York that has already been in use for several years now in Alabama, including Montgomery and Birmingham. This process looks at where crime has been in recent hours, days, months or even years, allowing police agencies to pick up on patterns that they might not otherwise recognize. It maps all of these various crimes, giving police officials the ability to collaborate on what resources might be most effective where.
PredPol, however, was developed in Los Angeles and works a bit differently. Developers boast that it predicts twice as much crime as existing systems of analysis.
The model will gather information about where crimes have occurred and then provide an updated model with red boxes to show where crime is likely to occur, say in the next 10 to 12 hours.
The computer systems don’t predict who is going to commit future crimes, but it does give officers a sense of where it is most likely to happen.
Using this information, police agencies assign patrol officers to the “red boxes” at various points in their shift, bolstering police presence in that area and encouraging whatever proactive measure officers are able to complete during that time. Those might include simple visibility, but it could also mean contacts with businesses, stopping suspicious persons in the area and just generally keeping an eye out.
For example, let’s say a certain area has a lot of parking lots in a given area. PredPol might use this fact plus past data to predict that the number of car thefts are going to be high in a certain area. Officers will then be told to be on the lookout for that particular kind of activity.
So far, it would seem there is nothing about these systems that could trigger legal challenges. But there are aspects about these programs that concern defense lawyers.
For example, there is that whole pesky business about something called reasonable suspicion.
An officer is required by law to have reasonable suspicion that a crime has been committed, is being committed or is about to be committed before he or she is legally allowed to stop someone. (It’s a lower threshold than probable cause, which is required for searches, warrants and arrests.)
So for example, just being near vehicles in an area where PredPol predicts a high likelihood of car jackings. Is it enough to create reasonable suspicion? There is some debate among officers, but our defense attorneys would unequivocally answer: No. You would have to be doing something else that would garner police suspicion.
Police making a stop based solely on one’s presence in these locations would almost certainly find that evidence suppressed in court.
Some departments have instructed officers not to use PredPol information as a basis for stops. However, if it is actually part of the foundation upon which an officer is basing his or her suspicion, one wonders how long they can realistically or legally deny that. This is especially true if the officer is put on the stand to testify.
There are some who would say that a computer isn’t biased, and therefore we may be able to eliminate some of the racial disparities apparent in average police stops. However, so long as the statistics feeding those computers do have the potential to be skewed, we will continue to see that reflected in future arrests as well.