A significant problem facing today’s prisons and jails is overcrowding. One of the biggest ways that the courts try to control the prison population is at bail hearings. This is when the crime the defendant is accused of is heard by the judge, and the state argues that the defendant should not be released on bail. The defendant, of course, would rather be released, so that he or she can continue to work, see their children or family, and live at home until their guilt can ultimately be assessed by the legal system. Judges are only human, and are prone to err. Many jurisdictions have thus started relying on the use of a pretrial computer system which uses algorithms to help determine if defendants should be released or held until they can be tried.

There are clear advantages to using this system: the decisions are often consistent, and leave little room for any judicial bias. The judges can use “validated, evidence-based data” when they release a defendant, so they can be confident of their decision. And of course, releasing more defendants means less jail crowding. But, the system is not perfect, and critics are now using the recent San Francisco tragedy of Edward French to argue against its use. The septuagenarian was out early one morning trying to catch the sunrise with his camera when he was allegedly shot and killed by two young men – Lamonte Mims and Fantasy Decuir. Mims had already been on felony probation after being arrested the year before and spent three months in jail, and again arrested this summer for gun possession and parole violations. He was released using this program and under the auspices of “assertive case management,” which requires frequent check-ins. The judge used the public-safety score issued by the computer to determine whether the defendant is a flight risk, or if he will return to court.

The system was originally developed by former Enron trader and Texas billionaire John Arnold, who runs a foundation aimed at criminal justice reform. Currently, twenty-nine counties use the tool, which examines nine risk factors such as age, criminal history, current and past charges. The tool then gives a number for the judge to consider. It was on this basis that the judge in San Francisco released a two-time felon who had violate parole, only to be arrested for the murder of another man a week later. According to the head of the non-profit that runs the tool, called the Pretrial Diversion Project, the score was a result of data which had been mistakenly entered into the program about Mims, giving him far fewer days spent in jail, resulting in a lower score. The problem is, the judge had access to the same data which was mistakenly entered into the program, with the correct numbers. Some fear that judges rely too heavily on these scores, even if their instincts and own interpretation of the information available conflict. Advocates of the tool simply say it is another tool that the judges can use when making their ultimate decision.

Regardless, most agree that the bail system needs reform, to change from a system where often 40 percent of defendants spend their time in jail awaiting trial simply because they cannot afford to pay the fine. Using this cash only system inevitably further punishes lower-income families rather than the more affluent, which smacks of inequity and unfairness. While the tool may not be effective, judges must use their discretion when making bail determinations, and many agree that the tool is a positive step forward in making changes to our bail system.