Opinion: PA’s new sentencing algorithm is a weapon of mass incarceration

The automated tool would increase racial bias and should not be implemented, says Frontline Dads’ Reuben Jones.

prison-kayleetornay-03
Kaylee Tornay / Billy Penn
Reuben (headshot)

Ed note: This opinion piece does not necessarily reflect the views of our news organization.


What did Walt Whitman mean when he wrote “I am large, I contain multitudes”? The famous words have many interpretations, and one of them is expressed in this quote from Bryan Stevenson, the groundbreaking civil rights lawyer: “Each of us is more than the worst thing we’ve ever done.”

Pennsylvania’s criminal justice system should look at people like me as containing those multitudes — instead of just as a statistic. Right now, the Pa. Commission on Sentencing is getting ready to do just that: to turn people convicted of crimes into ones and zeroes.

The commission was given a legislative mandate to build an algorithm to help judges sentence convicted people. The algorithm is designed to estimate who will be convicted again of a crime within three years of release. As the Philadelphia Defenders Association wrote for PennLive last week, the original purpose of the tool was “to assist judges in identifying those “with the lowest probability of being re-convicted of a serious crime.”

In other words, this tool was supposed to identify individuals at risk of receiving unduly harsh sentences.

But instead of developing an algorithm to send more people safely home, saving their communities money and distress, the commission has built one that could massively increase incarceration in the state. It’s also likely to exacerbate racial disparities that are already stark. In the Pennsylvania prison system, there are nine Black men for every white one.

Here I am in my multitudes. My name is Reuben Jones. I am the Executive Director of Frontline Dads, Inc. and the Philadelphia campaign coordinator for JustLeadershipUSA. I am a father, husband, mentor, poet and social justice advocate. I served on the Mayor’s Transition Team in Philadelphia and on the Universal Pre-K Commission, and I earned a Presidential Service Award from President Barack Obama. For the last 16 years I have worked diligently to promote equity and fairness for human beings who have been marginalized and who have not enjoyed the full rights and privileges they deserve.

But in 1987, I was convicted of robbery and aggravated assault and sentenced to 15 years in prison. I was 22 years old.

Because I was accused of a violent crime, I was viewed as a risk for inflicting harm, and given a lengthy sentence. I had a high bail of $150,000 because I was viewed as a flight risk. Because of my age, I was viewed as a person who would pose a risk of recidivism. I was not viewed as a person with potential or value. I was certainly not viewed — by the system, anyway, back then — as a person worthy of serving on a mayoral transition team or deserving of a presidential award.

I am still on parole from this conviction today. If arrested again, an algorithm would see me as a high risk. It would be wrong. In fact, even the commission thinks the tool can only predict high risk for conviction 52% of the time. But I would be called risky nonetheless and the tool would communicate that opinion to a judge with power over my future.

The tool before the commission heavily weights conviction record and age in its predictions — in a world where plea bargains are the norm — and are far harsher for us, for Black, Brown and poor people. I have a serious conviction, yes. But instead of being “risky,” I have been home for 16 years without any infractions or violations. Imagine if I was still in prison today. Imagine if I had not had the opportunities to redeem myself and prove to be an asset to my city.

Philly has lost thousands of Black men to prison, and over 3 percent of us are on probation or parole. Do we really want to increase that ratio with risk assessment? While the tool does not explicitly use race as a factor, it is trained to believe that Black defendants are riskier than white defendants — and the tool communicates that opinion to judges. If put into place, judges will equate risk with danger, and will lean more towards lengthier sentences to account for the high risk, or perceived dangerousness, of Black defendants.

The Pennsylvania Commission on Sentencing must not adopt the tool it is considering. Instead, I ask Governor Wolf and the General Assembly to continue investigating the racial bias of risk assessment tools; to only apply tools of this nature to send far more people home; and to guarantee more investigation into our humanity for everyone impacted by harm and mass incarceration, not less.

And I ask all of you who share my concern about Minority Report-style risk assessment to testify against it with me at the Criminal Justice Center, at 10 a.m. this Wednesday, Dec. 12, in Courtroom 504.

If we divest decision-making to an algorithm that ignores our capacity to change, then we will enter a new, frightening stage of mass incarceration. If the tool is trained to find risk, it will most certainly find it, and weaponize it to incarcerate us, rather than send us home to make amends, and to heal our families and communities.

Thanks for reading another Billy Penn story

Seems you’re the kind of person who really digs in. Want more? Sign up for our morning update, the quick, easy, free way to stay on top of Philly news.

Thanks for reading Billy Penn

Like the story above, everything we publish is powered by our members. If you enjoy reading, throw us a bone! Just $5/month makes more difference than you’d think.

Thanks for reading! We need you.

Member donations power our newsroom. If Billy Penn helps you feel more connected to Philly, we’d love to count you in. Will you join us?

Lock in your support

Reader support powers our newsroom. A monthly membership helps lock it in.

Can we count on you as a Billy Penn sustainer?

Winning the local journalism game

Thank you: Your support as a member is powering our news gathering.

Know someone else who might like our work? Invite them to join.