レポート 2415

The Chicago Police Department’s “Strategic Subject List” was quietly put out to pasture in November, according to the Office of the Inspector General, which had raised several issues regarding the program’s efficacy.

After 10 years, the Chicago Police Department has decommissioned an enormous database that, ostensibly, used analytics in an effort to predict which people are most likely to be involved in shootings — as either the shooter or the victim.

The “Strategic Subject List” was quietly put out to pasture in November, according to the Office of the Inspector General, which had raised several issues regarding the program’s efficacy. 

CPD used the list to identify which prior arrestees would be most likely to carry out shootings, or be victims of shootings — referred to as being a “party to violence.”

The list also was used to target gang members and their associates through information gathering, analysis and social network mapping, according to the office of Inspector General Joe Ferguson.

In a statement last week to announce the program’s demise, Ferguson’s office said: “The general areas of concern in the PTV [party to violence] risk model program include: the unreliability of risk scores and tiers; improperly trained sworn personnel; a lack of controls for internal and external access; interventions influenced by PTV risk models which may have attached negative consequences to arrests that did not result in convictions; and a lack of a long-term plan to sustain the PTV models.” 

Every person arrested by the CPD between August 2012 and June 2018 was assigned a “risk score” whether or not they were ultimately convicted of a crime. By July 2018, nearly 400,000 people had been assigned risk scores, according to the inspector general. 

The list was a largely secretive endeavor until 2017, when the CPD released a version of the list after the Sun-Times filed a lawsuit. 

The demise of the Strategic Subject List came shortly before the departure of Jonathan Lewin, the CPD’s former chief of technical services. 

“While these particular risk models have now been decommissioned, there are critically important and widely applicable lessons to learn here about the importance of careful data handling and thoughtful, purpose-driven policies and training, in ongoing and future efforts to harness data that predicts and ultimately prevents violent crime,” Ferguson said in a statement.