Algorithms predict our commute time, the weather, and, in an increasing number of states, whether a defendant should be released before trial. In the trending wave of bail reform currently underway across the nation, various states and local jurisdictions are abandoning cash bail systems in favor of technology that assists judges in determining whether to release or detain defendants by predicting their risk of rearrest or risk of failing to appear for court.[1] Although such risk assessment tools are a step in the right direction, skepticism remains. [2] What follows is a brief introduction to the most prominent risk assessment tool and an explanation of the criticism surrounding the use of such technology in general.
Funded by Arnold Ventures, the Public Safety Assessment (PSA) is a risk assessment tool currently used across the country and, of note, in large cities including Phoenix, Chicago, and Houston.[3] The PSA was created with pretrial records from 750,000 cases spanning over 300 jurisdictions, the largest and most diverse set of records ever assembled.[4] The PSA assesses nine factors, including a defendant’s age, prior convictions, pending charges, and prior failures to appear in court.[5] Based on these nine factors, the PSA generates three predictive scores: a “failure to appear” score, a “new criminal activity” score, and a “new violent criminal activity” score.[6] These scores are then combined to generate a raw score that determines a defendant’s placement on a six-point risk scale, from low to high risk.[7] However, the PSA stops short of directing judges to either release or detain a defendant pursuant to the defendant’s risk score.[8] Instead, jurisdictions have developed their own policy frameworks and guidelines to assist judges in implementing PSA scores when making pretrial release decisions.[9]
Critics of risk assessment tools maintain that the data on which these tools rely is neither neutral nor objective.[10] Because minorities have historically been disproportionately targeted by law enforcement, the data used to develop risk assessment tools is not devoid of racial bias; therefore, neither are their results, regardless of the fact that these systems do not consider a defendant’s race or demographics.[11] Just last year, more than 100 civil rights organizations, including the ACLU and the NAACP, signed a statement highlighting concerns with pretrial risk-assessment tools and urging those jurisdictions that have adopted them to regularly ensure that the systems are actually reducing jail populations and racial disparities.[12]
Additional criticism of risk assessment tools stems from fear that these systems will ultimately replace a judge’s own experience and discretion in making pretrial decisions.[13] In response to this, advocates for risk assessment maintain that these tools only serve to quantify determinations that judges are already making on their own when presented with the issue of whether to detain a defendant.[14]
On their own, risk assessment tools will neither ensure lower rates of detainment nor will they ensure pretrial systems without biases and inconsistencies. However, by shifting away from a pretrial decision-making process focused primarily on a defendant’s financial resources, risk assessment tools such as the PSA ensure that defendants are only detained when necessary to ensure their reappearance in court and the safety of the community.[15] Adopting a system “in which judges have access to scientific, objective risk assessment tools could further our central goals of increasing public safety, reducing crime, and making the most effective, fair, and efficient use of public resources.”[16]
*J.D. Candidate, 2020, Florida International University College of Law.
[1] See, e.g., Public Safety Assessment, https://www.psapretrial.org/about (last visited Oct. 25, 2019).[2] See Bryan Lynn, Judges Now Using Artificial Intelligence to Rule on Prisoners, VOA Learning English (Feb. 7, 2018), https://learningenglish.voanews.com/a/ai-used-by-judges-to-rule-on-prisoners/4236134.html.
[3] About the PSA, Public Safety Assessment, https://www.psapretrial.org/about#jurisdictions-united-states (last visited Oct. 25, 2019).
[4] See Public Safety Assessment FAQs (“PSA 101”), Arnold Ventures (Mar. 18, 2019), https://craftmediabucket.s3.amazonaws.com/uploads/Public-Safety-Assessment-101_190319_140124.pdf.
[5] About the PSA: Risk Factors and Formula, Public Safety Assessment, https://www.psapretrial.org/about/factors (last visited Oct. 25, 2019). “The nine factors do not include questions about race, gender, ethnic background, income, substance abuse, mental health, employment status, marital status, or any demographic or personal information other than age.” Arnold Ventures, supra note 4.
[6] Public Safety Assessment, supra note 5.
[7] Id.
[8] Id.
[9] Two such guidelines, the Decision Framework and Release Conditions Matrix, reflect local statutes, court rules, and policy preferences.“The Decision Framework is designed to support justice professionals in making decisions that align with the law, and the Release Conditions Matrix is structured to help judicial officers associate the PSA scores with the appropriate supports that may benefit a person while on pretrial release.” Id.
[10] See, e.g., Note, Bail Reform and Risk Assessment: The Cautionary Tale of Federal Sentencing, 131 Harv. L. Rev. 1125, 1132 (2018).
[11] Stephanie Wykstra, Bail Reform, Which Could Save Millions of Unconvicted People from Jail, Explained, Vox (Oct. 17, 2018), https://www.vox.com/future-perfect/2018/10/17/17955306/bail-reform-criminal-justice-inequality.
[12] More than 100 Civil Rights, Digital Justice, and Community-Based Organizations Raise Concerns About Pretrial Risk Assessment, The Leadership Conference on Civil & Human Rights (July 30, 2018), https://civilrights.org/2018/07/30/more-than-100-civil-rights-digital-justice-and-community-based-organizations-raise-concerns-about-pretrial-risk-assessment/.
[13] Lynn, supra note 2; Quentin L. Kopp, Replacing Judges with Computers is Risky, Harv. L. Rev. Blog (Feb. 20, 2018), https://blog.harvardlawreview.org/replacing-judges-with-computers-is-risky/.
[14] Wykstra, supra note 11.
[15] “In our society liberty is the norm, and detention prior to trial or without trial is the carefully limited exception.” United States v. Salerno, 481 U.S. 739, 755 (1987).
[16]Laura & John Arnold Found., Research Summary: Developing a National Model for Pretrial Risk Assessment 2 (2013).