We encounter algorithms all the time. There are algorithms that can guess within a fraction of a percentage point whether you’ll like a certain movie on Netflix, a post on Facebook, or a link in a Google search. But Risk Assessment Tools now being adopted by criminal justice systems all across the country - from Arizona, to Kentucky, to Pennsylvania, to New Jersey - are made to guess whether you’re likely to flee the jurisdiction of your trial, or commit a crime again if you are released. With stakes as high as this — human freedom — some are asking for greater caution and scrutiny regarding the tools being developed. Chris Bavitz, managing director of the Cyberlaw Clinic at Harvard Law School, helped draft an open letter to the state legislature of Massachusetts about Risk Assessment Tools, co-signed by a dozen researchers working on the Ethics and Governance of Artificial Intelligence. He spoke with Gretchen Weber about why we need more transparency and scrutiny in the adoption of these tools. Read the open letter here: https://cyber.harvard.edu/publications/2017/11/openletter