Machine learning principles can be applied to the internet of things and smart devices, like the Google car or fitness and medical wearables. And related risk prediction covers multiple sectors, including finance, healthcare, government, transportation and even the criminal justice system.
Machine learning could prove more efficient than judges’ predictions. Researchers from Cornell, Stanford, Harvard, and the University of Chicago, in collaboration with the National Bureau of Economic Research, studied “how machine learning can be used to improve and understand human decision-making” and developed an algorithm that can detect whether a suspect is a flight risk.
By analyzing large volumes of data such as police records, older cases and the defendant’s rap sheet, the algorithm has proven accurate and has so far aced prediction tests in New York.
The software, based on pattern recognition, will help judges predict defendants’ behavior (try to escape or commit more crimes). What’s more, it will reduce the costs of jailing a suspect during trial if not considered a flight risk and it “could cut crime and reduce racial disparities amongst prisoners.”
“We only observe crime outcomes for released defendants, not for those judges detained. This makes it hard to evaluate counterfactual decision rules based on algorithmic predictions,” the paper reads. “Second, judges may have a broader set of preferences than the single variable that the algorithm focuses on; for instance, judges may care about racial inequities or about specific crimes (such as violent crimes) rather than just overall crime risk.”
The project is still young and might need more thorough research to make sure the outcome is fair, and not a software error.