The point (as made clear in the article) is that a Human Being's decisions can at least, in principle, be challenged: "So you think the defendant poses a high

The point (as made clear in the article) is that a Human Being's decisions can at least, in principle, be challenged: "So you think the defendant poses a high risk of violence, recidivism or pretrial risk -- why exactly?" Not so with the "decisions" that were made by the proprietary algorithm described in the article.