Criminal justice advocates touted the end of cash bail as a way to remedy the fact that poor people had to stay in jail while the affluent went free. While the idea was good in theory, now that many states have abolished cash bail, it turns out that another factor is keeping the poor behind bars after an arrest.
That factor is the increased use of risk assessment algorithms, and it's leading to more incarceration, not less.
Risk Assessment Algorithms
While judges once determined whether or not a person received bail, now it's just as likely that an algorithm will make the decision. Of course, a computer algorithm is only as good as the information programmed into it, and many of these algorithms appear biased or reliant on faulty data.
California was the first state to eliminate cash bail, but the risk assessment program designed to replace it has made matters worse for many defendants. The program depends on algorithms to decide who is a good candidate for release and who should stay behind bars, but the system also allows local officials wide latitude on who is high risk and who is not.
Each individual receives a personal risk score that is supposed to foretell how likely an individual is to appear in court if released and the likelihood of their re-offending. The downside is that the personal risk score software is designed to compare people with similar profiles, with data retrieved from a criminal justice system and that criminal justice system has built-in discrimination.
Critics allege that even the most sophisticated software will replicate this discrimination and could make the situation worse for the people at the algorithm's mercy.
Replacing Bail with Profiling
California eradicated cash bail only to replace it with profiling, and it's not the only state. A 2017 study published in the Minnesota Law Review (although dealing with the Kentucky criminal justice system) showed that what advocates thought would occur when risk assessment went into place and what actually happened are poles apart.
In 2011, Kentucky implemented risk assessment as a mandatory part of the bail decision and was lauded as being on the cutting edge of reducing incarceration. However, an analysis of 1 million cases led to the study's conclusion that, while risk assessment led to a substantial change in the practice of setting bail, the number of those granted pretrial release didn't change significantly.
In addition, failure to appear and instances of pretrial crime increased. There was also another disturbing result of implementing risk assessment in Kentucky – the number of white defendants released increased while there was no change for black defendants.
Needs Assessment vs. Risk Assessments
Going back to a cash bail system isn't the answer, but what is? Perhaps replacing risk assessment with needs assessment, focusing on an individual's need for rehabilitation and other services. At the least, the risk assessment algorithms require a thorough overhaul, so they more accurately reflect the population without biases.
As it is, a defendant who is a person of color living in a high crime community may end up staying in jail because the algorithm finds they live in an area with a lot of former convicts. That fact – shared by thousands of defendants, many of them innocent – makes them a public safety threat as far as the risk assessment is concerned.
It's appalling, and it has to change. We can do better than this.