Highlights
Development of the US Department of Justice risk and needs assessment could affect as many as 106,000 federal prisoners.
If researchers and critics cannot come to an understanding as to what truly works, the future of criminal justice reform hangs in the balance.
The Congressional Budget Office estimated that the implementation of the First Step Act (via a risk instrument) would reduce federal prisoners by more than one-third or about 53,000 offenders. If states adopt it, thousands more could be released.
Author
Leonard Adam Sipes, Jr.Retired federal senior spokesperson. Thirty-five years of award-winning public relations for national and state criminal justice agencies. Interviewed multiple times by every national news outlet. Former Senior Specialist for Crime Prevention for the Department of Justice’s clearinghouse. Former Director of Information Services, National Crime Prevention Council. Former Adjunct Associate Professor of criminology and public affairs-University of Maryland, University College. Former advisor to presidential and gubernatorial campaigns. Former advisor to the “McGruff-Take a Bite Out of Crime” national media campaign. Certificate of Advanced Study-Johns Hopkins University. Aspiring drummer.
Currently, federal prisons are overcrowded. Good management requires judgments about people that are fair and free of bias
The Numbers
Out of the 6,741,000 people under daily correctional supervision, 870,000 are on parole compared to 3,790,000 on probation. 1,527,000 are in prison, Correctional Populations in the US.
The vast majority of released prisoners (five in six) will be arrested again. Over fifty percent of inmates will return to prison, Offender Recidivism.
Choices
We have to choose who gets prison and who gets intensive supervision on parole and probation. We want to focus on those posing an obvious risk to public safety with incarceration, supervision, and programs. The rest are labeled an acceptable risk (while overlooking a lot of criminality and bad behavior) for whom little is done.
It’s impossible to incarcerate everyone and not all offenders need programs and services. To put it bluntly, there is limited money and the justice system can’t afford the same levels of interaction with everyone.
Thus moderate to low risk people on parole and probation who have an endless number of violations or failures to pay restitution, fines and court costs “successfully” complete community supervision daily. Holding them accountable would break the bank and lead to administrative firings.
We within the justice system need to figure out who the truly dangerous people are. The future of the justice system depends on it. But it’s not simple.
Algorithms
Algorithms predicting personal risk have been around for decades via insurance companies. Efforts on the part of advertisers and polling companies seek to understand who we are, our opinions, and what we are willing to do or buy.
People seem to have an adverse reaction to algorithms (see Pew below) as something sinister offered by Google, Facebook and social media companies. But Google is doing nothing more than polling companies hired by General Motors did in the past as they try to understand who you are and what you are willing to do. Yes, Google and the others are doing it on a grand scale, but what did you expect for the right to use free search engines and email services?
Algorithms and artificial intelligence are being used to predict crime in cities, judge the dangerousness of criminal offenders, and sift through mountains of data regarding DNA, license plate readers and criminal investigations, National Institute of Justice.
But the big news is that the Department of Justice is trying to create a risk instrument to implement the First Step Act. Keep this in mind because, in the coming months, all hell will break loose regarding the controversy that a federal risk instrument will create.
Critics of risk instruments for criminal offenders offer two observations; they are biased instruments that fail to be any more accurate than flipping a coin.
Without risk instruments, criminal justice reform and the management of the offender population cannot proceed.
MuckRock
Per an article in MuckRock (mostly direct quotes, edited for brevity); Two months after First Step Act, no sign work has started on its centerpiece risk assessment software.
A bipartisan effort to address and reduce mass incarceration, the law increases the number of days of prison time waived for good behavior and proposes expansion of access to rehabilitative services.
Some inmates have already been released due to the First Step Act’s expanded compassionate release and revised sentencing provisions. But according to U.S. Sentencing Commission estimates, those changes will impact far fewer people than the eventual development of the risk and needs assessment, which could affect as many as 106,000 prisoners.
Within 180 days of enactment, originally June 19th, 2019, a tool would have been developed to make core assessments, including:
- classification, made at intake, of each federal prisoner’s risk of recidivism
- an assessment of each prisoner’s risk of violent or serious misconduct
- recommendations for the type, amount, and frequency of evidence-based recidivism reduction programs prescribed to each prisoner
There are currently more than 180,000 people in federal custody, and the CBO estimated that implementation of the First Step Act would by 2028 reduce that by more than one-third or about 53,000 prisoners.
Civil rights groups – 119 so far – have signed onto a letter of concern against risk assessment tools, emphasizing the importance of careful development and transparency about situations in which they are being used, MuckRock.
Pew, Algorithms, And Criminals
Per Pew (edited quotes), algorithms are all around us, using massive stores of data and complex analytics to make decisions with often significant impacts on humans – from choosing the content people see on social media to judging whether a person is a good credit risk or job candidate.
Pew Research Center released several reports in 2018 that explored the role and meaning of algorithms in people’s lives today. Here are some of the key themes that emerged from that research.
The public is wary of computer algorithms being used to make decisions with real-world consequences. The public expresses widespread concern about companies and other institutions using computer algorithms in situations with potential impacts on people’s lives.
More than half (56%) of U.S. adults think it is unacceptable to use automated criminal risk scores when evaluating people who are up for parole. When asked to elaborate about their worries, many feel that these programs violate people’s privacy, are unfair, or simply will not work as well as decisions made by humans, Pew.
Conclusions
The development of a risk instrument by the Department of Justice is a pivotal moment in criminology and criminal justice management. There are existing risk instruments within the Federal Bureau of Prisons and federal parole and probation agencies, but something as sweeping as this could have implications for the entire justice system.
It’s possible that criminal justice reform will either rise or fall based on this development. States and counties could adopt the federal risk instrument rather than develop one of their own; it provides them with cover and lets the feds take the inevitable heat and costs. If states adopt it, thousands more could be released from crowded state facilities.
The Federal Sentencing Commission has long argued the seriousness of the crime, criminal histories and the age of the defendant are the primary factors as to predicting future criminality, Offender Recidivism. But those who believe in risk instruments feel strongly that there are endless additional mitigating factors at play that “could” impact future criminality.
The seriousness of the crime, criminal history and age might be the “best” predictive factors, but education, support of family and friends, a willingness to engage in a job or substance abuse training may increase a person’s chances for change. What we did at twenty may not necessarily predictive of who we are at thirty.
But critics of risk instruments suggest bias throughout the process to the point where they are demanding the demise of algorithms to predict future behavior. Detractors state that risk instruments for offenders may not be useful predictors.
If researchers and critics cannot come to an understanding as to what truly works as to risk, the future of criminal justice reform could hang in the balance.
Contact
Contact us at leonardsipes@gmail.com.
My book: A “Best Business Book,” Success With The Media: Everything You Need To Survive Reporters and Your Organization available at Amazon
This is an ad-free website.
Reviews are appreciated.