Chicago's Strategic Subject List — A Decade of Predictive Policing Without Evidence It Reduced Violence

Starting in August 2012, the Chicago Police Department assigned every person it arrested an algorithmic "Strategic Subject List" (SSL) score estimating their risk of being a party to gun violence, as either victim or offender. By July 2018 nearly 400,000 people had been scored. The tool remained confidential until the Chicago Sun-Times obtained a version through litigation in 2017.

A quasi-experimental evaluation by RAND Corporation researchers (Saunders, Hunt and Hollywood, published in the Journal of Experimental Criminology) found that an early version of the SSL "does not appear to have been successful in reducing gun violence," though people on the list were more likely to be arrested — including for offenses that never led to conviction.

In January 2020, Chicago's Office of Inspector General (OIG) issued a public advisory citing the "unreliability of risk scores and tiers," inadequate staff training, weak access controls, and the absence of any long-term sustainability plan. The Chicago Police Department had already quietly retired the SSL the previous November, after roughly a decade of use. Inspector General Joe Ferguson called it a lesson in "the importance of careful data handling and thoughtful, purpose-driven policies."

The SSL is a cautionary rather than a success case: it shows a major police department deploying a predictive-risk algorithm city-wide for years before independent research and its own watchdog established that it was not achieving its stated public-safety purpose.

Read the full analysis: https://chicago.suntimes.com/city-hall/2020/1/27/21084030/chicago-police-strategic-subject-list-party-to-violence-inspector-general-joe-ferguson