Quantifying Criminal Procedure: How to Unlock the Potential of Big Data in Our Criminal Justice System

Abstract

Big data’s predictive algorithms have the potential to revolutionize the criminal justice system. They can make far more accurate determinations of reasonable suspicion and probable cause, thus increasing both the efficiency and the fairness of the system, since fewer innocent people will be stopped and searched. However, three significant obstacles remain before the criminal justice system can formally use predictive algorithms to help make these determinations. First, we need to ensure that neither the algorithms nor the data that they use are basing their decisions on improper factors, such as the race of the suspect. Second, under Fourth Amendment law, individualized suspicion is an essential element of reasonable suspicion or probable cause. This means that either the predictive algorithms must be designed to take individualized suspicion into account, or the predictive algorithms can only be used as one factor in determining whether the legal standard has been met, forcing police and judges to combine the algorithm’s results with individualized factors. And finally, the legal standards themselves must be quantified so that police and judges can use the numerical predictions of big data in their reasonable suspicion and probable cause determinations. These obstacles are not insurmountable. And if the necessary changes are made, the criminal justice system will become far more transparent, since the factors the algorithms take into consideration will necessarily be open for judges and the general public alike. Furthermore, setting a quantified likelihood for reasonable suspicion and probable cause will allow us to engage in a healthy debate about what those numbers ought to be, and it will also ensure conformity across different jurisdictions.