With Artificial Intelligence gaining quick popularity and evolving as we speak, many believe this is a “today” subject. However, in the late 90’s a software sold to state governments and police departments began tearing lives apart, in a way not even an attorney can defy.
A digital AI algorithm called COMPAS was created in 1998 by a company called Northpointe, Inc. (which now goes by the name equivant). This AI software is pretty much designed to predict the future. When someone is arrested, they fill out a long questionnaire about their life. The program combines their answers with their criminal record and compares them to thousands of past offenders. It then spits out a score from 1 to 10, 1 meaning the person is low risk (let them go home), and 10 meaning high risk (keep them locked up).
Since this software is owned by a private company, the exact mathematical formula used to calculate a person’s score is a proprietary trade secret. Which means that if you get a 9 out of 10, your defense attorney is legally banned from seeing how the computer came up with that number. You are judged by a formula that is hidden from the public view, and even from the one person that is supposed to make your case.
We would think that something judges trust so much would be impartial, however studies have proven this is not the case. An investigation by ProPublica in 2016 proved that the math is broken. The AI routinely made mistakes by flagging Black defendants as “high risk”, at twice the rate of white defendants, even when the Black defendants had cleaner records. Meanwhile, it mistakenly labeled white defendants as “low risk” far more often.
People that don’t deserve it have been denied bail, given longer prison sentences or denied parole based on a number they can’t even see the background of. Lives have been affected drastically by a software that is still flawed even after the research findings, and no one is doing anything to stop it. Racism is still taking freedom away from innocent people, because judges rather believe a number from a software than a story from a real human. A number from a secret software that has clear racist patterns should not be able to decide whether a Black father can see his kids sooner than later. This type of issue should not be happening in today’s age. We all bleed red, we all deserve fair judgement, and AI marking people as “high risk” due to skin color is not fair judgement.
When did we decide to trust a computer more than a human’s testimony? And why, after clear results, has justice not been served to the people done wrong by this AI system? People today keep being affected by it, and it keeps ruining lives. Will it ever change the way it judges people? Or will America’s justice system one day realize that a number from a racist computer is everything but fair judgement to their people?
For more articles like this, subscribe for free to Crucial Evidence down below.
This has been your author and expert witness.
Your girl,
Jimena Valle
No comments:
Post a Comment