Should We Be Suspicious About Predictive Policing?

|

The future of policing is now. It's not Robocop or drone surveillance but it is run by computers. New software called PredPol (for "predictive policing") is being used by as many as 20 of the nation's 50 largest police forces to target potential criminals in supposedly high-crime areas.

However, a new PredPol program in Los Angeles highlights the need for balance between crime prevention techniques of police and the citizen's protections from both racial profiling and unreasonable search and seizure.

The PredPol software uses an algorithm to indicate an area in which a crime is likely to occur that day. Police then station themselves in the area. The hope is their anticipatory presence will either deter a crime from happening or give them the ability to catch a criminal in the act.

What about discretion?

While the software seems beneficial for police, it may be bad for minority groups and disenfranchised citizens in the locations identified by PredPol. Too heavy of a reliance on software could cause police to consider PredPol's algorithm as reasonable suspicion that a crime has occurred or is about to occur in the area.

"Reasonable suspicion" is the legal standard needed by police to conduct searches, seizures and arrests. It is the same legal standard that is used in a "stop-and-frisk" detention, also known as a Terry stop. Proponents of so-called proactive policing say it often uncovers weapons and other "on the body" items that could be used to carry out a crime.

More broken windows?

Although a Terry stop is constitutional, as determined by the Supreme Court in 1968. The technique remains controversial, especially in large cities like New York where the "broken windows" theory on criminology was born. (New York police suspended use of the strategy after a U.S. District Court judge ruled in 2013 that the way they employed it violated individuals' constitutional rights.)

The "broken windows" theory states that police enforcement of minor criminal laws in urban areas helps to prevent more serious crimes over a greater period of time.

If police start looking for areas with high rates of vandalism and burglary, they could also use the data to pinpoint the location of more dangerous criminals - but it might not be that simple.

According to police watchdog organizations, the algorithm's predictions are only as good as the data used in the software. Bad data could have a compounding effect when put through an algorithm. If bad data is used to determine who might commit a crime or where, it could lead to unnecessary scrutiny of otherwise innocent citizens.

Categories: