To: Rep. McKell, M.,
Subject: End Secret Profiling in Criminal Justice... STOP Black Box Algorithms
Date: 2020-02-28T06:17:39Z
Body:
Dear Rep. Acton, There is a huge push to use algorithms to solve our criminal justice problems. Proponents and proprietors of such risk assessment tools think they can use computers to replace common sense and judicial discretion, all while preserving such abstract but important concepts such as justice. Indeed, it is believed that we can tell the police where to go arrest and who to arrest; we can decide who gets pretrial release or not; and then we can decide who gets what sentence and who is offered a second chance - all by algorithm. Yet, there are serious cracks in the computerized justice system, and it is time for states and jurisdictions using such risky algorithms to take a serious look at whether the use of these algorithms work, what impact they have on bias in the system, and evaluate the negative impact it may be having on outcomes. When it comes to getting out on bail pending trial, states and jurisdictions are moving to require an algorithm to evaluate criminal defendants to decide if they get bail or not, and if they do, whether they'll have their blood chemistry monitored, be on a GPS unit, or be subject to house arrest. Yet, there has been widespread criticism that such algorithms are inherently biased. In addition, algorithms that lack full transparency and are protected trade secrets are being used in jurisdictions all over the country preventing judges, defense lawyers, and prosecutors from seeing behind the curtain. In one landmark study of the use of such algorithms, they didn't decrease mass incarceration at all, only having a "trivial" decrease in the jail population. In sentencing, algorithms have been shown to be biased against African-Americans. In many states, these algorithms define who is suitable to be in the community and thus the algorithms' results can directly influence whether a person gets prison or probation. Sincerely, William Morgan American Fork, UT