On December 28, 2018, a petition was filed by Nevada Chief Justice James Hardesty to adopt a statewide requirement that all judges in Nevada utilize a validated risk assessment tool for use in pretrial decision-making. The petition was based on the recommendation by the Committee to Study Evidence-Based Pretrial Release.
The American Bail Coalition opposes this recommendation.
A public hearing on the proposal will be held on February 5, 2019 -- ADKT 0539: In re Adoption of Risk Assessment Tool for Use in Pretrial Decision-Making.
The American Bail Coalition submitted the following public comment regarding the petition and proposal to adopt pretrial risk tools:
Nevada Supreme Court
January 24, 2019
This comment is the response of the American Bail Coalition to the call for public comment in response to the petition filed by Justice Hardesty requesting that the Supreme Court approved his request to require all judges within Nevada to use a validated pretrial risk assessment tool for use in deciding bail and conditions of release from jail pending trial.
First, the Nevada Judiciary should not be in the business of approving a particular risk assessment algorithm or tool or requiring that judges use a particular tool. This creates an impermissible appearance of impropriety, calling into question the objectivity of the Court, because when the tool is ultimately legally challenged the Court of last resort in Nevada will have voluntarily put itself in a position to have been the body mandating by judicial edict the use of the tool and also having pre-endorsed the scientific validity and neutrality of the tool. Judges in Nevada can already consider risk assessment results, and the proponent of a risk assessment should be the party tasked with defending the scientific validity of the same.
Second, one-hundred ten national civil rights groups have called for an end to the use of pretrial risk assessments in the United States of America primarily due to concerns of racial bias and lack of transparency. Here, the validation study did not test for protected-class bias. Will the tool magnify racial and gender bias in Nevada? There is no way to tell. Further, the data is not open-source. I have asked for it twice, and my request has been denied. The information could be redacted so that I and other interested parties could conduct our own testing of it. But, it is truly a black-box system, one that was been rejected by the Leadership Conference on Civil Rights, the AI Now Institute, numerous other scholars, and interest groups from around the nation.
“Our communities need policies and practices that reverse mass incarceration, not ones that reinforce the racism already painfully present in the system. Until we address the inherent racism in our justice system and focus on meeting the needs of our communities, technological ‘solutions’ like risk assessments will continue to fall short.” – Scott Roberts, senior criminal justice campaign director at Color of Change
Third, this particular risk assessment under consideration relies heavily on demographic factors, such as employment, residential stability and age. Former USAG Eric Holder cautioned that using demographic factors could have the exact opposite impact as intended:
“Although these measures were crafted with the best of intentions, I am concerned that they may inadvertently undermine our efforts to ensure individualized and equal justice. By basing sentencing decisions on static factors and immutable characteristics – like the defendant’s education level, socioeconomic background, or neighborhood – they may exacerbate unwarranted and unjust disparities that are already far too common in our criminal justice system and in our society.”
Fourth, the risk assessment relies more heavily on un-convicted conduct than any risk other assessment I have seen. This is hugely problematic from the perspective of defendants’ rights. For example, three of the nine categories can be scored based on having been arrested but not convicted of misdemeanors, gross misdemeanors or felonies. In addition, if someone is homeless, they will get two points for that, labeling them as high risk solely due to their poverty. Moreover, a non-resident but legal alien would get a higher risk score due to the fact that they are not a “resident” of the State of Nevada.
At the end of the day, a jury could exonerate someone of all charges, and a defendant could sue civilly to seek redress for a false allegation and be awarded judgment, and yet this risk assessment will place them into the higher risk category, unnecessarily trammeling their civil liberties by labeling them as dangerous, a label that no subsequent exoneration would remove under this risk assessment. To add insult to injury, there is no explanation of the risk assessment will consider prior arrests that have been subsequently sealed.
Fifth, these tools don’t work and have been shown to increase failures to appear in court as required, increase crimes while on bail, and have virtually no impact on the reduction of the pretrial population (and in some cases dramatically increasing the pretrial population). A landmark study by Professor Stevenson from the Anton Scalia School of Law demonstrated this quite clearly, which article is in the materials.
“The primary conclusion is that risk assessment in Kentucky came “not with a bang but a whimper.” 17 Despite being crafted with the explicit goal of lowering incarceration rates,18 HB 463 led to only a trivial increase in pretrial release. Furthermore, the increase in the release rate was matched by an uptick in failures-to-appear (FTAs) and pretrial crime; a disappointing counter to hopes that all three margins could be improved simultaneously.” – Megan Stevenson, Assessing Risk Assessment In Action
Sixth, at least one scholar blames the generational increases in mass incarceration on the use of risk assessment tools and labeling people as dangerous which has occurred and exploded in America’s penal system since 1970. This article is also in the materials. Said Professor Werth of Rice University:
“It has been argued that risk assessment tools could help stem the tide of mass incarceration,” Werth said. “However, the evidence suggests that thus far, risk assessment instruments have contributed to expanding the number of people enmeshed in the criminal justice system – encompassing imprisonment, probation and parole.”
Seventh, State v. Loomis, a case in which the U.S. Supreme Court denied certiorari, the Wisconsin Supreme Court held that without adequate due process warnings regarding the risk assessment tool in that case, which is quite similar to the tool under consideration here, that the use of the tool would violate the due process clause. In the case of the Nevada tool, there is no are no such limitations or restrictions on use.
We are attaching an appendix of materials that we believe the Supreme Court should review prior to making a decision in this matter. While risk assessment procedures were indeed in vogue at the time the committee was launched by Justice Hardesty, the issue has pivoted so substantially that revisiting the issue once again in light of the substantial quantity of scholarly work that has been done in the last two years, some of which we have put in the appendix, is the only appropriate way forward.
In approving a risk assessment that has not been tested for racial or other protected-class bias and relies so heavily on demographic factors and unconvicted conduct, the Supreme Court of Nevada is disregarding a known risk that the tool may disparately treat similarly situated defendants. For that reason alone, the Supreme Court should decline to require the use of this risk assessment.
 881 N.W.2d 749 (Wis. 2016)(“Although we ultimately conclude that a COMPAS risk assessment can be used at sentencing, we do so by circumscribing its use. Importantly, we address how it can be used and what limitations and cautions a circuit court must observe in order to avoid potential due process violations.”
 For example, one tool that is similarly constructed as the Nevada tool, was biased against African-American defendants: link to ProPublica story and link to thesis from Dartmouth U.
More than 100 Civil rights, Digital Justice, and Community-Based Organizations Raise Concerns About Pretrial Risk Assessment
Assessing Risk Assessment in Action – Megan Stevenson
Theorizing the Performative Effects of Penal Risk Technologies: (Re)producing the Subject Who Must Be Dangerous – Professor Robert Werth
Study: Risk assessment tools may increase incarceration rates
Detention, Release From Jail, and Computerized Bail Justice in California: Is it 1984 All Over Again? What Can California Learn From the Last 30 Years of Bail Reform? – UCLA Criminal Justice Law Review
The Biased Algorithm: Evidence of Disparate Impact on Hispanics
Dartmouth Study – Accuracy of Risk Assessments
Algorithmic Transparency for the Smart City
Machine Bias: There’s software used across the county to predict future criminals. And it’s biased against blacks – ProPublica
“Fair” Risk Assessments: A Precarious Approach for Criminal Justice Reform
The Third Generation of Bail Reform in America: Is the Third Time the Charm – Jeff Clayton, @LAW Winter Edition
The 4th Generation of Bail Reform – American Bail Coalition
Comment on Proposed Rule Changes, Addition of Rules 4.10 and 4.40 of the California Rules of Court