Risk Assessments Put Your Safety At Risk

Senator John Kennedy, who represents Louisiana in the U.S. Senate, recently wrote an opinion piece for Fox News titled, “Bail, bond decisions are being made today with algorithms – that puts your safety at risk.”

He stated that, “Algorithms diminish public safety in this country. They ask us to pretend that lengthy arrest records and violent crimes don’t matter.  They ask police to scoop up the bad guys only for the courts to immediately release them.  They turn us into a bad joke.”

Risk assessments have been marketed as the panacea for bail reform all over the country. Risk assessments are designed as a systematic process for evaluating the potential risks of a defendant to be involved in a future criminal activity or pose a danger to the community. They also determine if a defendant is likely to appear for court. A risk score is assigned based on such evaluation and recommends if the defendant can safely be released into the community while their case is winding through the criminal justice system.

Risk assessments supposedly take the bias out of the release equation and doesn’t depend on whether or not someone has the financial means to secure their release. In essence, a fairer pretrial release system. Violent offenders would not be recommended for release unless strong conditions are added, while those charged with non-violent offenses would most likely be recommended for release. This is what those who propose the use of risk assessments would have you believe.

Fact: all across the country dangerous and violent offenders have been given very low assessment risk scores, which in turn translates into a recommendation for release.

Senator Kennedy in his opinion stated that the Metropolitan Crime Commission found:

37.6 percent of people arrested for violent felonies in New Orleans during the third and fourth
quarters of 2018 received the lowest risk level score of one; and

That included more than 32 percent of the people arrested for homicide and 36.5 percent of people arrested for rape.

Specific examples of defendants released by a low risk assessment score in Louisiana include:

Alan Parson and Richard Sansbury were charged with the robbery of a New Orleans pharmacy in which workers were tied up and a police officer shot. They received a risk assessment score of “one.”

Neither Parson nor Sansbury is from Louisiana. They are believed to be part of an Indianapolis gang responsible for pharmacy robberies in multiple states. But a computer algorithm deems them a minimal risk to public safety.

Theron Glover, age 18, currently faces 160 criminal charges in New Orleans, many related to car break-ins and auto theft. He received a risk assessment of “three” after a series of heists.

While out on bail, Glover is suspected of progressing from stealing Honda CRVs to firing into a crowd of people. Four people were injured, and Glover was hauled back to court. This time, the algorithm gave him a risk assessment score of “one.”

In fact, a broad coalition of more than 100 civil rights, digital justice and community-based organizations released a share statement highlighting their concerns of using algorithmic-based tools for release decisions. They also stressed that such tools can worsen, not help, racial disparities and allow for further incarceration.

Release decisions should only be made by a judge who is elected by the citizenry and who will be held accountable for their release decisions. Algorithmic-based risk assessment tools have no such accountability. They take the human equation out of the release decision without having the full facts on a defendant’s criminal history and any failures to appear.

Victims of crime deserve justice and accountability in the criminal justice system.