Since 2012, scoring algorithms created to manage risks in the United States penal system have been adopted by Immigration and Customs Enforcement (ICE) agencies across the United States. First developed as a mechanism to reduce anti-immigrant biases and to ensure the humane treatment of detainees, scoring algorithms suffered constant revisions to accommodate DHS enforcement priorities as well as the preferences and punitive biases of ICE agents. With the arrival of the Trump administration, a technology created to ensure the humane treatment of undocumented immigrants became central to the policy of criminalization of the immigration process. This book provides historical, qualitative, and quantitative evidence of the process that placed risk assessment technologies at the forefront of the anti-immigration battle.

Using very large data sets on immigration and detention proceedings obtained from DHS through multiple FOIA requests, this Brief reveals the inner workings of the risk classification algorithms (RCA) used to process tens of thousands of immigrants each day. Chapters examine the tension between risk algorithms and end users and explain how ICE officers' preferences shape the scoring properties of RCA used in immigration enforcement. Illustrating how scoring algorithms oppress immigrants of color, this book is interest policymakers, immigration scholars, lawyers, criminologists, political scientists, and university professors, graduate students, and undergraduate students in the behavioral sciences.

Analyzes immigration detention using risk classification assessment data Introduces empirical analysis of immigration detention using Oaxaca-Blinder decomposition method Shows how the Trump Administration optimized punitive bias in immigration detention

Autorentext

Robert Koulish (University of Wisconsin, Ph.D.) is the Director of MLaw Programs, administered through the College of Behavioral and Social Sciences (BSOS) at the University of Maryland. Koulish is also Research Professor in Government and Politics and Lecturer at Law at the UMD Carey School of Law in Baltimore, where he teaches immigration law and policy. He is co-founder of the Crimmigration Control International Net of Researchers (CINETS, 2011), and author of Immigration and American Democracy (Routledge, 2009), co editor of Crimmigrant Nations (Fordham U. Press, 2020), Immigration Detention, Risk and Human Rights (Springer 2016), Crimmigration in the Age of Covid-19 (Social Sciences, 2023), and several articles including "The Immigration Detention Risk Assessment" (with Mark Noferi, 2014), and "Manipulating Risk" (with Kate Evans, 2020). He has researched Central Europe and along the U.S. Mexico border, investigating human rights abuses, refugees, and patterns and practices of social control within legal institutions.

Ernesto Calvo (PhD, Northwestern University 2001), Professor of Government and Politics at the University of Maryland, Director of the Interdisciplinary Lab for Computational Social Science (iLCSS), and Field Editor of the Journal of Politics (JOP-Comparative). His work uses big data to study comparative political institutions, political representation, and social networks. He is the author of Non-Policy Politics (with Victoria Murillo, Cambridge, 2019), Legislator Success in Fragmented Congresses in Argentina (Cambridge, 2014), and over 50 publications in Latin America, Europe, and the United States. The American Political Science Association has recognized his research with the Lawrence Longley Award, the Leubbert Best Article Award, and the Michael Wallerstein Award. He is currently working on a book project on activating political content in social media.

Inhalt

Preface.- Technologies of Crimmigration.- A theory of punitive biases in Risk Assessment Algorithms.- From Humane to Criminal Detention Changes in RCA Scoring Rules.- Punitive Bias and Algorithm Updating Measuring Puni-tive biases through officers dissents.- Optimized at Last Making RCA Criminals out of Undocumented Immigrants.- Concluding Remarks The Crimmigration Calculus.