Predicting crime: 'Powerful tool' needs policy
The Rochester police use intelligence-led policing to identify people who perpetrate a high volume of crime in the city. But that is only half of a greater tool that could, eventually, also identify and predict when and where crime occurs.
This new technology, called SPSS — Statistical Package for the Social Sciences — is the new topic for policy discussion between the Rochester Police Department and the Police Policy Oversight Commission.
Police Chief Roger Peterson described SPSS as a technology with great potential for predictive analytics, but also one with almost no existing policies for use by local governments.
"We've discussed potential uses for it, and there are many," Peterson told commissioners on Monday. "It can be an extremely powerful tool, but like any powerful tool, it can also be misused. We want to have that discussion before we start applying it."
SPSS works by mapping crime, when and where it occurs and under what circumstances, Peterson said. The potential application would be to manage the police department's resources and to place officers in the right places at the right times.
There is also a great concern, Peterson said, that the technology would be viewed by the public as "high-tech profiling" if the tool led to police deploying more resources in minority neighborhoods.
"It becomes a potential for high-tech profiling, and that's obviously what we want to prevent it being," he said.
The police department already owns the equipment and had been part of a pilot program, Peterson said. It is not being used currently and would not be used without discussion of a policy governing the tool.
Peterson suggested police department staff present the technology during the commission's next meeting. The discussion would reflect, "Here's what it can do; what do you want it to do?" Peterson said.
"I like how you're suggesting we approach it," said Commission Member Don Barlow. "This can be a great win for the community."
Barlow also emphasized the importance of understanding the technology and its predictive potential before diving into policy making.
"Unless we have an understanding as to what the predictive modeling is capable of, then we're just kind of hearing information," Barlow said.
The commission meets next on Dec. 14 at 6 p.m. In the city-county Government Center, Room 104.