Predictive Policing: A Potential Nightmare for Communities of Color


Predictive policing may be a dream come true for budget-strapped police departments, but the implications of its use in youth and communities of color are a nightmare.

“Predictive policing” refers to the use of data analysis software to guide law enforcement; it employs data on crime and offenses to generate predictions on the places and times that future crimes may occur. The software developers promise lower crime rates without the investment of additional staff or money.

Predictive policing companies have been gaining popularity across the country, most recently with Oakland Mayor Libby Schaaf’s potential new investment of over $150,000 in PredPol Inc.’s software.

Yet, despite the bold claims made by  software developers, there has been no independent study conducted to confirm the effectiveness of predictive policing. Additionally, the negative implications of these methods for youth and communities of color have been grossly overlooked by law enforcement and completely ignored by software developers.

Crime forecasting can only be as good as its data. At the fundamental level, predictive policing is destined for failure.

The W. Haywood Burns Institute relies on law enforcement data to determine where inequalities exist within the system and to inform policy changes. We’ve discovered, and countless independent studies support, that there are vast racial and ethnic disparities throughout youth and criminal justice systems across the country. As long as these disparities exist, the implementation of predictive policing software in communities of color will lead to the continuation, if not worsening, of racial and ethnic disparities.

To use a simplified example –

Communities of color are policed at higher rates than predominantly White communities. Therefore, certain crimes that go unnoticed in White communities (e.g. marijuana possession) more often result in criminal charges in communities of color, despite similar usage across all demographic groups.

Predictive policing software documents arrests made, not the actual crimes committed. Because communities of color tend to experience vast disparities in arrest rates when compared to their White peers, those neighborhoods are shown to be “hot spots.” Thus, a cycle of over-policing in communities of color and the resulting racial and ethnic disparities in justice systems, continues to persist.

While the collection and analysis of data is critical to the justice system, the statistical methods used by predictive policing confuse correlation and causation to the detriment of the lives of people of color and their and communities.

Predictive policing methods should raise concerns for all communities. As a RAND report released in 2013 stated, “The very act of labeling areas and people as worthy of further attention inherently raises concerns about civil liberties and privacy rights.”

While Oakland may be attracted to the simplicity of fighting crime promised by PredPol, Inc., the decision to invest thousands into an unproven program is questionable at best. For a police department still under federal control for repeated abuses of power, the adoption of such a controversial program could have disastrous effects for communities of color throughout Oakland.