Predictive policing, the trendy field in law enforcement that utilizes data collection and analysis to try to predict where crimes will occur, has been very popular among police departments in the United States.
Unfortunately, this technology has not always been proven to work well. In fact, a new study, published this week, seems to suggest that, in at least one case. It’s been a total waste of time… and money.
A new joint investigation by The Markup and Wired shows that, for the city of Plainfield, New Jersey, predictive policing has been a gigantic and expensive mess, producing almost no useful results.
Spending money on useless software? Yes, please
Both outlets analyzed a large dataset provided by the Plainfield police, which included around 23,631 predictions made by Geolitica, a crime prediction software.
Geolitica (formerly known as PredPol but changed its name two years ago) claims to use “data-driven strategies” to help police identify so-called “hotspots,” meaning the places where crimes are more likely to occur.
However, the dataset provided by the Plainfield Police revealed that (over a period of approximately ten months, between February 25 and December 18, 2018) the software accurately predicted where crimes would occur with an almost 0% success rate.
Gizmodo and The Markup previously collaborated on an investigation into PredPol software, discovering that police were disproportionately targeting low-income communities of color using the technology.
The English outlet attempted to contact Geolitica to inquire about the study’s findings but was unsuccessful. The company is slated to cease its operations by the end of this year; some of its team members have already been hired by other companies.