Tom Muscarello, a computer scientist at Chicago’s DePaul University, has been working since the mid-1990s on perfecting an artificial intelligence system aimed at helping the Chicago Police Department blaze a new trail in the way it solves serial robberies, rapes, and other violent crimes, the Chicago Tribune reports. The computer system, called the Classification System for Serial Criminal Patterns (CSSCP), is expected to begin live trials at the Chicago Police Department as soon as early next year. The system uses pattern-recognition software that acts like a superhuman brain.
It will be able to cull massive amounts of data, pulling out details of individual crimes, such as the assailant’s age, sex, height, location of the crime, weapons and vehicle used, to create a criminal profile that can be compared with others. The goal is to help overcome the fact that detectives can have difficulty linking serial cases. The system also will have the potential to search through words or phrases in police reports, such as a criminal who wears “green military fatigues” or one who says “give it up” every time he robs a bank. It can work 24 hours a day without human intervention, sorting through thousands of criminal records per second, revealing patterns in seemingly unrelated crimes that a mere mortal could miss. “It could revolutionize the way [the Chicago Police Department] does police work,” said Charles Padgurskis, former director of information systems for the Chicago police, who has worked with Muscarello since the research project was launched a decade ago. In a study using three years of Chicago police robbery data (not active cases), the computer system — which uses a computer network particularly suited for this type of inquiry called a Kohonen neural network — detected at least 10 times as many related crimes as a team of detectives with access to the same data.