Once again, reality is trying to catch up with fiction. Any concept similar to Minority Report doesn't require mutant humans, but rather a huge amount of data. In the past, we came across Blue CRUSH, a true pioneer in systems for crime prevention. Now, we find PredPol, a platform used in 60 American cities, which hopes to expand to more metropolitan areas and demonstrate how effective the software is at predicting crimes.
Tensions Between Police and Citizens
In recent months, the relationship between American citizens and their police has deteriorated drastically. The cases of Eric Garner and Michael Brown raised the alarm about a notorious increase in police violence, which in turn has led to lethal reprisals, as happened in December 2014 with the death of two police officers in New York. At the end of January, a public defender was arrested for protecting her client (that is, doing her job), who was also African-American, once again sparking the issue of racism (Garner and Brown were both African-American). In other words, police forces are now fighting to restore trust. They not only have to do their job correctly, but also need to be more efficient. And when it comes to efficiency, nothing beats software covering their backs.
How PredPol Works
PredPol is the product of a company based in California that has been on the market for three years, and is also one of the first to adopt crime prediction as a specialization. PredPol works using three central data points: Type, place, and time of the crime, which are processed with the help of a proprietary algorithm, and additional historical details about the analyzed region. The result is a series of red boxes, between ten and twenty, that serve as recommendations for patrol routes. In its first year using PredPol, the Santa Cruz Police Department reported a 9 percent reduction in assaults, an 11 percent in illegal entries, and a 27 percent in armed robberies. Added to this are the numbers from the Richmond Police Department, which recorded a 21 percent decrease in violent crimes, 28 percent in property crimes, 50 percent in residential burglaries, and 34 percent in vehicle thefts.
Scrutiny and Costs
Even so, PredPol is still under the microscope. It still has to undergo more advanced studies, while other similar systems revealed no benefit. At the same time, there is concern that the red boxes are automatically interpreted as danger zones, which could lead to more unnecessary violence. Another focus of concern is an overconfidence in Big Data, causing serious errors if its functioning is not completely understood. Nevertheless, police departments are spending six-figure sums per year to access PredPol's predictions, and everything seems to indicate that the investment will accelerate.
