AI is helping Chicago police predict where crimes will happen next

It sounds like something out of a dystopian sci-fi novel, but it is happening now: artificial intelligence is helping Chicago police predict where crimes will happen before they take place. The technology is being applied specifically to Chicago's dangerous South Side, where police are given predictions about crime locations via a computer algorithm and digital maps. The same technology is being tested in cities around the nation, including Tacoma, WA, Denver, CO, and Lincoln, NE.

The use of technology by Chicago PD has been spurred by the South Side's massive murder rate, though overall crime rates have dropped since the technology went live. According to Reuters, the city's January to July shooting rates decreased 39% year-on-year, while murder rates have dropped a single percentage point to 33%. Districts utilizing this technology have also seen massive decreases in shootings ranging up to 29%, as well as up to 18% fewer homicides.

The technology works in part using a control room located in Chicago's 7th District police department. Operators monitor predicted crime locations on digital maps with the algorithm doing the work. Multiple technologies are being tested, such as Azavea's HunchLab with the 7th District. With this, existing crime data is correlated with relevant local data on everything from an area's socioeconomic status to which businesses are nearby and what the weather is like.

A total of eight elements are factored into the prediction process, though some of the finer details about the technologies being used are kept quiet. Though the overall prediction doesn't seem so unusual, what we do know about the technology is surprising. For starters, the computer system is assigning individual people a risk factor number that ranges from 0 (no risk) to 500 (maximum risk). Things like past drug arrests and gang affiliation are factored into a person's risk rating.

It is also known that a gunfire detection system is live in South Side Chicago (and other cities in the US), though how many sensors are being used isn't known. This system works to detect gun shots and the approximate location from which they originated, alerting police and helping them get there, hopefully, in time to apprehend the shooter. Such technology is necessary in a city where shootings — and murders — are happening in massive numbers.

The larger computer algorithm technology isn't without its critics — some fear it will pave the way for a dystopian world in which all people are known to law enforcement, monitored based on judged risk, and in some way subjected to punitive actions based on predictions about behavior. More immediate, though, are worries from some that reliance on this technology would cause a focus on stopping individual crimes at the expense of solving the larger problems from which they stem.

SOURCE: Reuters