Is it really 2054 already? That’s the year in which the Tom Cruise vehicle “Minority Report” takes place. In the movie, Cruise heads a futuristic crime-fighting unit that bases arrests and convictions on visions by three psychic beings who can see murders before they happen.

What was once the realm of science fiction has now become reality with the introduction of new crime prevention software in Washington D.C. — ironically the very city in which “Minority Report” is set. Developed by Richard Berk, Professor at the University of Pennsylvania, the software analyzes about two-dozen variables, including a person’s criminal history and their geographic location. The two most important variables in predicting future criminal behavior based on prior criminal acts are the type of crime and the age of the perpetrator.

The software is already being used in Philadelphia and Baltimore to predict which individuals on parole or probation are the most likely to murder and to be murdered. The newest version of the software that is being implemented in our nation’s capital, however, goes one step further, identifying individuals who are most likely to commit crimes other than murder.

If the tests in Washington are a success, Berk says the program could help set bail amounts and suggest sentencing recommendations.

Using software to predict human behavior is obviously controversial, especially when it is being used to make decisions that profoundly impact people’s lives. Besides being an action movie, “Minority report” was a cautionary tale about the dangers of trying to predict the nuances of human nature.

Fellow researchers have praised Berk’s algorithm for its accuracy, at least compared to previous attempts to develop murder-predicting software. Among high-risk groups the murder rate is one murder for every one-hundred people. Trying to predict such a rare event is very difficult. Using Berk’s new technology, however, researchers at UPenn were able to predict eight out of every one-hundred murders, rather than just one in one-hundred.

The benefits of effective crime-predicting software are obvious. By identifying high-risk individuals, the authorities should be able to better monitor those individuals and prevent more violent crimes in the future. But at what cost?

Indeed, the use of similar predictive analytic software, developed by IBM, was tested by The Florida State Department of Juvenile Justice and the UK Ministry of Justice in the spring and summer drew withering criticism.

Some worry about the possibility of false positives, and prejudging and punishing people for the crimes they could potentially commit. Others are concerned about potential abuses of the system and argue that the use of such technology violates the Constitution. Still others point out that the combination of such software with omnipresent surveillance could lead to the police scouring social-networking sites like Facebook for signs of potential criminal behavior.

Perhaps the most disturbing aspect of this technology, however, is that the youth of the offender is one of the best predictors of criminal behavior in the future. While the correlation between youth offenses and future offenses may be scientifically sound, it does not mesh with our societal concept of the clean slate. A child or young man who commits a crime may be statistically more likely to commit a crime in the future, but doesn’t he also deserve a second chance?

Jeremy Francis

Image Source

Comments are closed.