At the request of his probation officer, Tyrone C. Brown came to a community auditorium here in June and sat alongside about 30 other mostly young black men with criminal records — men who were being watched closely by the police, just as he was.
He expected to hear an admonition from law enforcement officials to help end violence in the community. But Mr. Brown, 29, got more than he had bargained for. A police captain presented a slide show featuring mug shots of people they were cracking down on. Up popped a picture of Mr. Brown linking him to a criminal group that had been implicated in a homicide.
“I was disturbed,” said Mr. Brown, who acknowledges having been involved in crime but denied that he had ever been involved in a killing.
That discomfort was just the reaction the authorities were after.
Mr. Brown, whose criminal record includes drug and assault charges, is at the center of an experiment taking place in dozens of police departments across the country, one in which the authorities have turned to complex computer algorithms to try to pinpoint the people most likely to be involved in future violent crimes — as either predator or prey. The goal is to do all they can to prevent the crime from happening.
The strategy, known as predictive policing, combines elements of traditional policing, like increased attention to crime “hot spots” and close monitoring of recent parolees. But it often also uses other data, including information about friendships, social media activity and drug use, to identify “hot people” and aid the authorities in forecasting crime.
The program here has been named the Kansas City No Violence Alliance, or KC NoVA. And the message on that June night to Mr. Brown and the others was simple: The next time they, or anyone in their crews, commit a violent act, the police will come after everyone in the group for whatever offense they can make stick, no matter how petty.
Such was the case for Mario Glenn, a 28-year-old with a criminal history that includes drug trafficking and assault. After he attended a program meeting, called a call-in, last year, he was caught during a police sting to take down a group implicated in several homicides. Mr. Glenn robbed a confidential informer trying to buy a gun from him, the police said. He has been convicted, and prosecutors are now seeking the maximum 30-year prison sentence.
“We have a moral reason to do a better job at addressing violence in this community,” said Jean Peters Baker, the prosecutor for Jackson County, which includes Kansas City. “I don’t know that this will work, but we need to try.”
The use of computer models by local law enforcement agencies to forecast crime is part of a larger trend by governments and corporations that are increasingly turning to predictive analytics and data mining in looking at behaviors. Typically financed by the federal government, the strategy is being used by dozens of police departments — including Los Angeles, Miami and Nashville — and district attorneys’ offices in Manhattan and Philadelphia.
At a time when many police departments are under fire for aggressive tactics, particularly in minority neighborhoods, advocates say predictive policing can help improve police-community relations by focusing on the people most likely to become involved in violent crime.
Civil liberties groups take a dim view of the strategy, questioning its legality and efficacy, and asserting that it may actually worsen the rapport between the police and civilians.
Ezekiel Edwards, the director of the Criminal Law Reform Project of the American Civil Liberties Union, said predictive policing tended to legitimize the profiling of racial minorities who live in poor, high-crime neighborhoods, and prompted officers to enforce laws selectively.
“Our concern is guilt by association,” Mr. Edwards said. “Because you live in a certain neighborhood or hang out with certain people, we are now going to be suspicious of you and treat you differently, not because you have committed a crime or because we have information that allows us to arrest you, but because our predictive tool shows us you might commit a crime at some point in the future.”
But researchers working with the police to develop the predictive algorithms say that they can come closer than traditional detective work to figuring out who is most apt to break the law. They say criminals commit violent crimes in fairly distinctive patterns and often have similar attributes. Those include previous arrests; unemployment; an unstable home life; friends and relatives who have been killed, are in prison or have gang ties; and problems with drugs or alcohol.
In Chicago, the police have developed a “heat list” of 400 people who are considered far more likely than the average person to be involved in violent crime. Factors in compiling that list included their criminal records, social circles and gang connections. Also a factor was whether they had been victims of an assault or a shooting.
“These are the people with the highest propensity for violence in the city of Chicago,” said Robert J. Tracy, chief of crime control strategies for the Chicago Police Department.
Read more: NY Times