Researchers from New York University (NYU) School of Law and NYU's AI Now Institute are suggesting that predictive policing systems may intensify discrimination in the criminal justice system by relying on so-called “dirty data,” according to a new study.

Because law enforcement has been heavily scrutinized in recent years for allegedly inordinate aggression toward minority suspects, experts have wondered whether technologies like predictive policing software — software used to help law enforcement identify potential criminal activity and its perpetrators based on data analytics — could potentially diminish instances of discrimination. However, NYU researchers concluded that such techniques potentially worsen discrimination when informed by dirty data that emerges from flawed or racially biased practices.

To reach that conclusion, researchers looked at data from case studies in Chicago, New Orleans and Arizona’s Maricopa County.

"We chose these sites because we found an overlap between extensively documented evidence of corrupt or unlawful police practices and significant interest, development, and current or prior use of predictive policing systems. This led us to examine the risks that one would influence the other," explained Jason Schultz, a professor of clinical law and one of the paper's co-authors.

The team pinpointed 13 jurisdictions (including those previously mentioned) with recorded instances of biased and unlawful police practices that simultaneously explored or deployed predictive policing techniques. For instance, researchers noted that the Chicago Police Department incorporated a system for identifying those at risk of either becoming a victim or a perpetrator of a homicide or a shooting while at the same time the department was under federal investigation for unlawful police practices. According to the study, the very same demographic of residents identified as targets of Chicago’s policing bias by the Department of Justice were essentially the same as those identified by Chicago’s predictive policing systems.

"In jurisdictions that have well-established histories of corrupt police practices, there is a substantial risk that data generated from such practices could corrupt predictive computational systems. In such circumstances, robust public oversight and accountability are essential," Schultz said.

The study, "Dirty Data, Bad Predictions: How Civil Rights Violations Impact Police Data, Predictive Policing Systems, and Justice," appears on SSRN.

To contact the author of this article, email mdonlon@globalspec.com