Prisons Turn to Computer Programs for Parole Decisions
As states continue to cut costs at correctional facilities, prison parole boards are turning to computer algorithms to help determine whether or not an inmate should be released. At least 15 states now require parole board judges to use some type of risk assessment method to aid in the decision, according to The Wall Street Journal. The tool mostly comes in the form of software and considers 50 to 100 different factors about a person before stating the likeliness that they would return to prison during a parole period.
The new system is working, too. The Wall Street Journal reported that the number of inmates in state and federal prisons fell almost 1 percent to 1.6 million in 2011, compared to just .01 percent the year before. However, those numbers don’t only reflect the fact that prisons are releasing more inmates. More inmates are actually staying out of prison. The Journal reported that about 12 percent of inmates who received parole were re-incarcerated in 2011, which is less than the 15 percent of parolees in 2006.
The computer systems determine whether or not a person should be released based on details such as an inmate’s age during their first arrest, whether the inmate believes the conviction to be fair or not and the inmate’s level of education. These assessments are designed to recognize patterns that can predict future crime. Some experts are concerned that these details can lead to racial bias because some of these factors can speak indirectly to a person’s race. This is part of the reason that parole boards will still continue to oversee and overrule risk assessments. Developer of the system Compasi told the Journal that its decisions are overruled between 8 percent and 15 percent of the time.
While the systems are not perfect, they help better select who should be released on parole and help with the cost of operations in prisons. Members of a parole board would also consider things like an inmate’s age and whether or not an inmate shows remorse for their crime; however, those don’t always align with how well they might do on parole. For instance, murderers and sex offenders are often less likely to commit another crime than someone guilty of a lesser crime, which a parole hearing wouldn’t necessarily determine on its own.