Predictive policing tools spread in the 2010s. Place-based systems such as PredPol, developed with the Los Angeles Police Department, flagged small areas for patrol. Person-based systems, such as Chicago's Strategic Subject List, scored individuals by their supposed risk of violence.
The core problem is a feedback loop. The data reflects where police have chosen to patrol and arrest, not where crime actually occurs. More patrols produce more arrests in the same places, which the software then reads as proof of more crime.
Critics, community groups and audits challenged these systems. Chicago shelved its list after an inspector general found problems with it. Los Angeles stopped using PredPol in April 2020 after an inspector general review found insufficient evidence that it reduced crime. Similar tools continue to operate in other cities, often with little public oversight or transparency.