On Sunday night, a select group of Seattle police officers working out the city's East and SW Precincts began their beats with an innovative new tool—one that forecasts where crimes are likely to occur based on the time, date, and type of crime committed within the city in the past five years.
The technology, called Predictive Policing software, was developed in Los Angeles and uses geographical data and crime statistics dating back to 2008 to create a complex algorithm that pinpoints likely crime hot spots on any given day and time, right down to 500-feet-by-500-feet, or a little larger than one square city block. Seattle Police Chief John Diaz and others explained that the hot spots are places that are generally well-known to officers as high crime areas already, but the software allows officers to devote roughly 30 percent of their day—the estimated amount of time that they're not responding to 911 calls—to actively patrol areas where criminal activity is most probable, given the time of day and history of related crimes in the area.
"The predictions are twice as effective as a human analyst using the same data," explained Mayor Mike McGinn at a press conference this morning. The LAPD was able to reduce property crime by 13 percent in a year using the software.
"It's very exciting," said Chief Diaz. "Neighborhoods are going to see this as an incredible tool."
For now, the East and SW Precinct are only piloting the software on property crimes, as those are the most common in our city. Eventually, the department hopes to have the technology up and running in every precinct and for every crime, especially to help "reduce gun crimes," explained Lt. Brian Grenon, who's helping with the Seattle pilot. The program will cost the city $73,000 for the software and a $45,000 annual subscription fee.
To be clear, we're still a few years away from arresting people for thoughtcrime. No information on past suspects is included in the algorithm—just the type of crime committed. When asked if the data forecast could be seen as another form of biased policing—by sending officers to patrol certain neighborhoods in South Seattle that have historically high crime, for instance—McGinn was insistent that the technology would, in fact, do the opposite. "This minimizes the influence that unconscious bias can have," he said, by relying on data and not on an officer's personal experience with known suspects in the area or their gut intuition.
McGinn acknowledged that "there are people with more power and influence in our community” that exert their influence to monopolize city resources. He added: “I hear from all communities that they want to be safe... this tool will help us get there."