Reasonably Suspicious Algorithms : Predictive Policing at the United States Border

As big data's promises of increased efficiency and serendipitous insights spread across a broad range of sectors, they are accompanied by new risks, some intuitive, some unpredictable. That dichotomy is heavily accentuated in the law enforcement context, where blithe application of new technologies to analog doctrines poses a greater threat to individual rights, and the potential of data analytics to add efficiency, accuracy, and accountability to existing procedures could be all the more beneficial. Predictive policing algorithms, which approximate the probability of crimes occurring in certain areas, or being committed by certain people, epitomize this dual dynamic, as they have the potential to increase accuracy and efficiency, but also threaten to dilute the reasonable suspicion standard and increase unintentional discrimination, in a manner that existing law is ill-equipped to prevent. This is of particular concern at the United States border, where Fourth Amendment protections are weaker due to the long-recognized governmental prerogative to investigate external threats poised to infiltrate the country. This Note will argue that while the use of predictive policing algorithms at the border should not be barred outright, so as to permit potentially beneficial uses of the technology to develop, it should be carefully limited by statute to prevent the wholesale trammeling of privacy and civil liberties