Justice in the Age of Big Data
Posted on March 18th, 2021 by Emilio Miles
In chapter 5 of Weapons of Math Destruction, Cathy O’Neil discusses the use of police models and their effect on their respective communities. While the models help to stop crime overall, they do so in a way that targets race as well as the poor and impoverished, despite claiming to overlook such aspects. The result is more and more data that helps support and sustain these flawed models.
O’Neil talks about how PredPol, a predictive program, was originally intended to help the police crack down on violent crime. This was due to the Reading police suffering from a huge cut in the number of officers after the 2008 market crash. Police chief William Heim had to figure out a more efficient way to control crime with fewer officers. Using PredPol to predict areas where crime was more likely to occur, chief Heim helped bring down the crime rate a year later.
Eventually, however, violent crimes were becoming less and less prominent so in order to meet their quotas, police started feeding the PredPol program with information on petty, non-violent crimes. PredPol didn’t go after “rich” crimes such as committing fraud or paying off rating agencies. Those types of crimes didn’t check off the boxes of the PredPol program. They required special tools and individuals that only institutions like the FBI and the SEC had, so they instead focused on the “mean streets.” This caused impoverished areas to suffer from the recidivism model that O’Neil talked about in earlier chapters. The data obtained from that, in turn, helped to reinforce the program’s validity. The problem with that becomes clear when people committing the same mistakes, such as public intoxication, in fraternities somehow end up getting off scot-free while the ghettos suffer.
So what can be done to prevent or improve upon this? One solution would be to use predictive programs such as PredPol to only target violent crimes. For petty, non-violent crimes police should focus their attention into getting closer and building trust with their communities. The idea is that if the goal becomes building trust, then an arrest might become a last resort instead of the first. This would prevent many of the tragedies that have come about in the recent years. I believe the biggest take away from this chapter is coming to understand just how far the reach of WMDs extends. In the end, though, a lot of the problems are ultimately due to human judgement and avarice.