Automating Justice: State Policy Variations, Discretion, and the Impact of "Clean Slate" Record Relief

Session Submission Type: Panel

Abstract

As state governments increasingly turn to automated policymaking and algorithmic adjudication to deliver public benefits, criminal justice reform has emerged as a primary frontier for "no-touch" administration. This panel explores the implementation, efficacy, and unintended consequences of "Clean Slate" automated record relief across California, Colorado, Michigan, Minnesota, Oklahoma, and Virginia. Clean Slate policies have garnered broad bipartisan support, as automating record relief expands access and reduces the burden on individuals to navigate complex expungement processes. Yet the full potential of these laws remains largely unrealized without meaningful follow-through, such as proactive notification, tangible reentry support, and pathways beyond simple relief. This panel asks: what should Clean Slate policies have accomplished, and what structural, administrative, and political barriers have stood in the way?

The four papers presented provide a multi-disciplinary examination of how Clean Slate policies operate in practice, with a specific focus on the intersection of automation, discretion, and equity.

The first paper examines the implementation of Clean Slate in Colorado, where prosecutors retain authority to object to cases identified for automated sealing. Findings show that while objections are relatively rare (6.6%), they are racially structured: Hispanic and Black individuals face significantly higher odds of objection relative to White individuals, suggesting that human discretion can reintroduce disparities even within automated systems. 

The second and third studies explore the experiences and needs of those receiving relief. Across five states, authors use original survey and interview data to theorize how individuals impacted by "no-touch" policies understand the role of the state. Findings show that while only 39% of respondents were aware of the automatic version of the policy, eligible individuals deeply value official state recognition and documentation as "proof" of rehabilitation. This highlights a tension between the efficiency of automation and the human need for state acknowledgement. The third paper focuses on a population that may see outsized benefits from from Clean Slate: Medicaid-eligible addiction treatment clients in Michigan. Using results from a pre-pilot survey of 80 individuals, the authors outline a strategy to adapt record expungement to the specific needs of this population, demonstrating how the removal of legal barriers can support retention in treatment and reduce the social stigma.

The final paper looks at the longer-term effects of record relief. Leveraging a natural experiment in California where 1.2 million people received record relief simultaneously, the study finds statistically significant but substantively small increases in arrest likelihood. The authors also document an implementation gap of several million people with conviction records remaining despite appearing statutorily eligible. This gap likely reflects the use of conservative algorithms to process relief in the face of incomplete administrative data.

Together, these papers form the basis for a critical discussion regarding the current state of Clean Slate policies in the U.S. By highlighting how implementation challenges can mute intended outcomes and affect who ultimately benefits, the panel assesses the potential and limitations of automated relief to improve socioeconomic outcomes and meaningfully reduce racial disparities.

Policy Area

• Crime, Justice, and Drugs

Chair

• Nefara D Riesch, Center for Policing Equity

Discussants

• Amanda Yvonne Agan, Cornell University

• Laura Chavez, The Clean Slate Initiative

Organizer

• Alissa Julia Skog, University of California, Berkeley

Individual Presentations

• Discretion Within Automation: Prosecutorial Objections Under Colorado’s Clean Slate Act - Presenting Author: Vivian Aranda-Hughes, Michigan State University

• Seeking State Recognition in Automated Policymaking: The Case of Clean Slate Criminal Record Expungement - Presenting Author: Joanna Wahmhoff, Northeastern University; Non-Presenting Co-Author: Sarah Lageson, Northeastern University; Non-Presenting Co-Author: Ericka B Adams, San José State University; Non-Presenting Co-Author: Elsa Chen, Santa Clara University

• Criminal Records Expungement Study - Presenting Author: Meghan M. O'Neil, Michigan State University

• Does automatic conviction relief affect recidivism: evidence from California - Presenting Author: Alissa Julia Skog, University of California, Berkeley; Non-Presenting Co-Author: Johanna Lacoe, University of California, Berkeley