Letter to Attorney General Barr RE: The use of the PATTERN risk assessment in prioritizing release in response to the COVID-19 pandemic
View this letter as a PDF here.
April 3, 2020
Honorable William P. Barr
U.S. Department of Justice
950 Pennsylvania Avenue, NW
Washington, DC 20530
RE: The use of the PATTERN risk assessment in prioritizing release in response to the COVID-19 pandemic
Dear Attorney General Barr:
On behalf of The Leadership Conference on Civil and Human Rights (The Leadership Conference), a coalition of more than 220 national organizations committed to promoting and protecting the civil and human rights of all persons in the United States, and the undersigned organizations and individuals, we write to express our grave concerns with your March 26, 2020 memo to the Federal Bureau of Prisons (“Bureau of Prisons” or “Bureau”), concerning the prioritization of home confinement in response to the COVID-19 pandemic. We ask that you rescind this memo in its entirety, and want to share particular concerns with the use of PATTERN — a risk assessment system built as a result of the First Step Act — as a factor in determining which currently incarcerated individuals may receive “priority treatment” in transfer and release decisions. Further, we ask that you immediately work to safely reduce the federal prison population using your broader existing authority, as well as the expanded authority afforded to you under the Coronavirus Aid, Relief, and Economic Security Act (CARES) Act.
COVID-19 is spreading rapidly in cramped U.S. prisons and jails, where hundreds of thousands of vulnerable people face sickness and death. Already, hundreds of prisoners and staff have tested positive, and the first federal prisoner died from COVID-19 on March 22, 2020. This is a health emergency that urgently requires expedited releases from incarceration to enable social distancing and to protect people in prison and jail, correctional staff, and communities.
On March 26, 2020, you issued a memorandum directing the Bureau of Prisons to transfer some vulnerable people from prisons to home confinement in the name of minimizing their exposure to COVID-19. We believe the restrictions you identify in this memo for home confinement eligibility are extremely troubling. In particular, you indicate the Bureau should rely upon an assessment tool, PATTERN, which numerous civil rights and legal organizations have previously warned is problematic and likely to perpetuate racial disparity in decision-making. Moreover, your directive to the Bureau regarding PATTERN’s use for home confinement decisions during an emergency health crisis was not its intended use, and limits transfer prioritization to those assessed as “minimum risk.” The use of a tool like PATTERN to make life or death decisions is alarming and serves to justify leaving tens of thousands of people — mainly people of color — unprotected and at the mercy of a deadly pandemic.
Your memo instructs that only people who receive a “minimum” risk score from the PATTERN tool will receive “priority treatment.” According to data furnished by the Department of Justice in February, individuals classified as minimum risk are the smallest cohort of the federal prison population, compared to the groups identified as low, medium, or high risk. Also, based on an analysis of PATTERN using a sample of the federal prison population, and reported by the National Institutes of Justice, only 7 percent of Black men in the sample were classified as minimum, compared to 30 percent of White men. This indicator alone should give the Department of Justice great pause in moving forward with the memo’s directive.
Experts have repeatedly criticized PATTERN, noting that the tool is scientifically unverified, and that the assumptions built into its design encode bias against Black people, Latino people, poor people, unhoused people, and people with mental illness.
Given the unprecedented and immediate risk that COVID-19 poses to people in prison and prison workers, we categorically reject the use of PATTERN or any other recidivism risk assessment tool to justify leaving vulnerable people incarcerated.
- The undersigned organizations and individuals recommend that the Department of Justice and Bureau of Prisons use their many authorities to decarcerate as many people as possible through all avenues of release, from all facilities under their control, as soon as possible. No release should be conditioned on electronic monitoring, nor should those afforded accelerated release be subject to heightened surveillance.
- The undersigned organizations call on the Department of Justice and Bureau of Prisons to abandon the use of PATTERN and all other recidivism risk assessment tools for any form of release recommendation or decision-making, now and in the future.
- PATTERN risk scores cannot inform assessment of medical risk and should not play any role in determining who receives access to adequate healthcare. As jail and prison protocols are developed to facilitate accelerated release and accommodate CDC safety guidelines for those incarcerated and facility personnel, punitive measures such as solitary confinement must not take the place of adequate health care. Moreover, institutional protocols must ensure that public health measures do not have collateral effects on the behavioral records—automated and otherwise—of those incarcerated.
Our Specific Concerns with PATTERN
- Predictions of recidivism are wholly inappropriate for informing medical release and public health in the context of a pandemic.
- Because PATTERN’s forecast of “general recidivism” is based on an incredibly broad definition of “re-offend”, it is highly likely to produce assessments that are biased against Black people and people of color, and that disproportionately impact those experiencing homelessness, or living with mental health issues.
- The memo says that the risks of detention must be weighed against the risks of release; however, the “general recidivism” score produced by PATTERN does not help in making this assessment.
- When tools conflate the likelihood of arrest for any reason with risk to public safety, a large number of people will be labeled a threat without sufficient justification. Risk assessments that include minor offenses or technical violations in their definition of “risk” will inflate risk scores and incarceration rates and exacerbate racial inequalities. In the context of COVID-19 and this memo, this means a much higher risk of illness and of fatality.
- Limiting release to people with a “minimum” risk score will produce significant racial bias
- By conditioning release decisions on PATTERN risk scores, the Bureau of Prisons is poised to leave Black people and people of color disproportionately exposed to harm. Under this memorandum, only 7 percent of Black men currently incarcerated would receive “priority treatment,” whereas 30 percent of White men would. This disparity is the byproduct of historical patterns in the ways different racial groups are treated differently by the criminal justice system.
- Black people and people of color are treated more harshly than similarly situated White people at each stage of the legal system, which results in serious distortions in the data upon which PATTERN relies. Historical court and arrest data primarily reflect the past and present operations of the criminal justice system, recording who police chose to arrest, how judges choose to rule, and which people are granted longer or more lenient sentences. By relying on this data, PATTERN systematically overestimates the risk of people of color. There are no technical fixes to these problems that could make PATTERN and similar tools safe and fair to use.
- These biases are compounded by the fact that people of color have increased risk of illness and death from COVID-19 infection, due to structural health inequalities.
- The choice of PATTERN’s risk thresholds — in other words, the process of determining how many individuals are rated minimum or low risk by PATTERN — was not done with this pandemic in mind. Under this memo, an individual who is assessed as having an 89 percent chance of “success” upon release (meaning they would not be rearrested or have a technical violation within 3 years) would not be prioritized for potential release.
- We are in an unprecedented situation, and historical arrest data are irrelevant to any assessment of public safety risk, especially during a pandemic.
- PATTERN’s validity rests on the assumption that criminal history data can serve as a reliable and neutral measure of underlying criminal activity, but such records cannot be relied upon for this purpose.
- Any predictions based on historical arrest data are ill-suited to make predictions about public safety risk in the current moment. Given the extraordinary circumstances under which we are currently living, historical crime data amount to “zombie” data — meaning the data used to build these models do not apply to our current conditions.
- Despite fears of increased criminal activity, local police are reporting that crime levels (including violent crime) have plummeted to some of the lowest levels seen in years. Thus, there is strong reason to believe that the likelihood of arrest for any crime would be much lower than historical patterns indicate.
- Moreover, part of the promise of PATTERN was that it would give people the opportunity to reduce their risk scores by participating in programming. Yet, not enough time has passed for people to take advantage of this opportunity. As a result, it is unlikely anyone has had the chance to meaningfully alter their risk score, let alone be reassessed.
In conclusion, our communities for years have warned decision-makers — including the Department of Justice — and the public about the risk of predictive technologies in high-stakes human decision-making systems. Tools like PATTERN are unfair, biased, and wrong on their own merits. But using them in a process to decide who gets the right to access social distance and freedom in the worst global pandemic in generations is particularly wrong. Therefore, we urge you to use your existing and expanded authority under the CARES Act to transfer as many people as possible into home confinement, without any of the limitations articulated in your memo, given that hundreds of thousands of lives are at stake. If you have any questions, please feel free to contact Sakira Cook, Director, Justice Reform Program, The Leadership Conference on Civil and Human Rights, at [email protected].
- 334 East 92nd Street Tenant Association
- A Little Piece of Light
- AI NOW
- Alabama Justice Initiative
- All of Us or None, Bakersfield
- Alliance of Families for Justice
- Alternate Roots
- American-Arab Anti-Discrimination Committee
- Autistic Women & Nonbinary Network
- Beauty After the Bars
- Believers Bail Out
- Bend the Arc
- Black and Pink Boston
- Block Builderz
- Buried Alive Project
- Carceral Tech Resistance Network
- California Coalition for Women Prisoners
- California Legal Research
- Campaign for Youth Justice
- Center for Disability Rights, Inc.
- Center for Justice Research – Texas Southern University
- Center for Law and Social Policy (CLASP)
- Center on Privacy & Technology at Georgetown Law
- Charles Hamilton Houston Institute for Race and Justice
- Church of Scientology National Affairs Office
- Coalition for Women Prisoners NYS
- College and Community Fellowship
- Community Justice Exchange
- Cornell University
- Council on American-Islamic Relations (CAIR)
- Criminal Justice Policy Program, Harvard Law School
- Criminal Justice Program, UCLA School of Law
- Critical Race Studies Program, UCLA School of Law
- CURE (Citizens United for Rehabilitation of Errants)
- Defender Impact Initiative
- Defending Rights & Dissent
- Dignity and Power Now
- Dream Deferred
- Dream Deferred Inc
- Drug Policy Alliance
- Entre Hermanos
- Equal Justice Under Law
- Equality California
- Essie Justice Group
- Fair and Just Prosecution
- Faith in Texas
- FAM Queen Team
- Families for Justice as Healing
- Florida Legal Services, Inc.
- Free Hearts
- Giving Others Dreams – G.O.D
- Harm Reduction Coalition
- Haverford College
- Health in Justice Action Lab, Northeastern University School of Law
- Human Rights Watch
- IBW 21st Century Police Accountability Task Force
- Innocence Project
- Jewish Council for Public Affairs
- Just Futures Law
- Justice For Housing
- Justice Strategies
- Justice Support Group
- Juvenile Law Center
- LatinoJustice PRLDEF
- Lawyers’ Committee for Civil Rights Under Law
- The Leadership Conference Education Fund
- The Leadership Conference on Civil and Human Rights
- Legal Action Center
- Legal Services for Prisoners with Children
- Life After Release
- Life for Pot
- Massachusetts Against Solitary Confinement
- Massachusetts Bail Fund
- Matters of the Heart
- Media Alliance
- Media Mobilizing Project
- NAACP Legal Defense and Educational Fund, Inc.
- National Action Network
- National Association of Criminal Defense Lawyers
- National Association of Social Workers
- National Bar Association
- National Council for Incarcerated & Formerly Incarcerated Women & Girls
- National Council of Churches
- National Disability Rights Network
- National Immigration Law Center
- National Lawyers Guild
- NETWORK Lobby for Catholic Social Justice
- New Beginnings Reentry Services, Inc.
- New Direction Coaching & Consulting, LLC
- New Haven Women’s Resettlement Working Group
- NYU Law Center on Race, Inequality, and the Law
- OVEC-Ohio Valley Environmental Coalition
- Participatory Defense Hubs
- People’s Paper Co-op
- Pillars of The Community Participatory Defense
- Presbyterian Church (USA) Office of Public Witness
- Pretrial Justice Institute
- Prison Policy Initiative
- Public Justice Center
- ReEntry Matters
- Reintegrated Voices
- Release Aging People in Prison/RAPP
- Reproductive Justice Inside
- Resilience OC
- Richmond Community Bail Fund
- Rise and Resist
- Robert F. Kennedy Human Rights
- T.O.P. – Surveillance Technology Oversight Project
- Silent Cry Inc.
- Silver State Equality-Nevada
- Southeast Asia Resource Action Center
- State Vs Us Magazine
- T’ruah: The Rabbinic Call for Human Rights
- Texas Civil Rights Project
- The Bail Project
- The Black Sex Worker Collective
- The Daniel Initiative
- The Decarceration Collective
- The Greenlining Institute
- The Healing Project
- The National Council for Incarcerated and Formerly Incarcerated Women and Girls
- The National Reentry Network for Returning Citizens
- The Tadini House
- The Talking Drum Incorporated
- The United Methodist Church – General Board of Church and Society
- Tucson Second Chance Community Bail Fund
- UCLA School of Law
- Upturn Toward Justice in Technology
- Union for Reform Judaism
- Union Theological Seminary
- United Methodist Women
- University of Chicago Law School
- Washington Lawyers’ Committee for Civil Rights and Urban Affairs
- What’s Next Washington
- Witness to Mass Incarceration
- Women Against Mass Incarceration
- Women on the Rise
- Women Who Never Give Up, Inc
- Working Families Party
- WV Citizens for Clean Elections
- Young Women’s Freedom Center
- Chelsea Barabas; Doctoral Candidate, MIT
- Ruha Benjamin, PhD; Associate Professor, Princeton University
- Meredith Broussard, PhD; Associate Professor, New York University
- Joy Buolamwini; Founder, Algorithmic Justice League
- Sasha Costanza-Chock, PhD; Associate Professor, MIT
- Kate Crawford, PhD; Distinguished Professor, Co-Founder, Co-Director, AI Now Institute, NYU
- Colin Doyle; Criminal Justice Policy Program, Harvard Law School
- Bernard E. Harcourt, PhD; Professor of Law & Political Science, Columbia University
- Stefan Helmreich, PhD; Professor & Elting E. Morison Chair, MIT
- Martha Minow; 300th Anniversary University Professor, Harvard University
- Cathy O’Neil, PhD; Author, Weapons of Math Destruction
- Rodrigo Ochigame; Doctoral Candidate, HASTS, MIT
- Heather Paxson, PhD; Professor of Anthropology, MIT
- Seth J. Prins, PhD MPH; Assistant Professor, Columbia University
- Vincent Southerland; Executive Director, Center on Race, Inequality, & the Law, NYU School of Law
- Meredith Whittaker, Co-Founder, AI Now Institute and Minderoo Research Professor, NYU
- Jordi Weinstock; Lecturer on Law, Harvard Law School
 Coronavirus Aid, Relief, and Economic Security Act available at https://www.congress.gov/116/bills/s3548/BILLS-116s3548is.pdf
 Office of the Attorney General, “Prioritization of Home Confinement As Appropriate in Response to COVID-19 Pandemic,” Memorandum for Director of the Bureau of Prisons, March 26, 2020.
 The Leadership Conference on Civil and Human Rights, Comment Letter to Department of Justice on PATTERN First Step Act, available at https://civilrights.org/resource/comment-letter-to-department-of-
 U.S. Department of Justice, The First Step Act of 2018: Risk and Needs Assessment System, available at https://nij.ojp.gov/sites/g/files/xyckuh171/files/media/document/the-first-step-act-of-2018-risk-and-needs-assessment-system_1.pdf
 We outline our specific concerns below.
 U.S. Department of Justice, The First Step Act of 2018: Risk and Needs Assessment System – UPDATE, January 2020, at 12. (“A return to BOP custody or a re-arrest within three years of release from BOP custody, excluding all traffic offenses except driving under the influence (DUI) and driving while intoxicated (DWI).”) See also Brandon L. Garrett, Megan T. Stevenson, Open Risk Assessment, Behav Sci Law. 2020; 1–8.
 U.S. Department of Justice, The First Step Act of 2018: Risk and Needs Assessment System, p. 62, Table 8 available at https://nij.ojp.gov/sites/g/files/xyckuh171/files/media/document/the-first-step-act-of-2018-risk-and-needs-assessment-system_1.pdf.
 Decades of research have shown that, for the same conduct, African-American and Latino people are more likely to be arrested, prosecuted, convicted and sentenced to harsher punishments than their white counterparts. For decades, communities of color have been arrested at higher rates than white communities, even for crimes that these racial groups engage in at comparable rates. Megan Stevenson & Sandra G. Mayson, The Scale of Misdemeanor Justice, 98 B.U. L. Rev. 731, 769-770 (2018). For example, African Americans are 83 percent more likely to be arrested for marijuana compared to whites at age 22 and 235% more likely to be arrested at age 27, in spite of similar marijuana usage rates across racial groups. Ojmarrh Mitchell & Michael S. Caudy, Examining Racial Disparities in Drug Arrests, Just. Q., Jan. 2013, at 22. Similarly, African-American drivers are three times as likely as whites to be searched during routine traffic stops, even though police officers generally have a lower “hit rate” for contraband when they search drivers of color. Ending Racial Profiling in America: Hearing Before the Subcomm. on the Constitution, Civil Rights and Human Rights of the Comm. on the Judiciary, 112th Cong. 8 (2012) (statement of David A. Harris). This leads to an overrepresentation of people of color in arrest data.
 Of particular concern is the requirement for a person to have a “minimum” risk score in order to be prioritized for release. In order to be assessed in the broad category of minimum risk, an individual must be assessed as “minimum” risk for both general and violent recidivism. This requirement will result in a large number of Black and Latino people being deprioritized for release, given historic racial disparities in arrest rates.
 See Simone Weichselbaum, Weihua Li, “As Coronavirus Surges, Crime Declines in Some Cities,” The Marshall Project, Mar. 27, 2020, available at https://www.themarshallproject.org/2020/03/27/as-coronavirus-surges-crime-declines-in-some-cities. (“In fact, in Chicago, Detroit, Los Angeles and San Francisco, recent data show big drops in crime reports, week over week. The declines are even more significant when we compare this year with the same time periods in the three previous years.”)
 According to the Department of Justice, as of January 15, 2020, every person currently incarcerated received an “initial” PATTERN score and was “assigned to participate in evidence-based recidivism reduction programs.” Less than three months have passed since that date — too little time for those incarcerated to have been reassessed based on the completion of programming.