Cohort 2025
Redo.io
Reshaping Sentencing of Low-Level, Nonviolent Offenses
Los Angeles County, CA
A growing body of research highlights the racial disparities in California’s criminal legal system, particularly those affecting people convicted of low-level, nonviolent offenses. In 2021, almost 20,000 people were incarcerated in California for nonviolent felonies alone, over 40 percent of whom were sentenced to six years or more. Black people are incarcerated at a rate approximately nine times greater than white people because of equal-protection violations and policies like the “Three Strikes and You’re Out” law, which disproportionately affected Black and Latinx populations through higher arrest and sentencing rates. Resentencing initiatives offer a crucial opportunity to identify and correct historical injustices owing to overly harsh sentencing.
With Catalyst grant funding, Redo.io aims to help identify incarcerated people serving long sentences for low-level, nonviolent offenses who may be eligible for second-look case review through its Resentencing Data Initiative. It aims to develop a robust web application using Microsoft tools that district attorneys and public defenders can use to directly interact with prison sentencing data, either using its own internal data or its public database of 95,000 California prison sentencing records. The application will include a scenario-based model to allow users to type out short queries for specific data and an AI assistant to help users explore the data, create visualizations, and run preliminary analyses. Redo.io will also compile and return cohorts of eligible candidates for attorneys to reference in their case reviews. Redo.io will partner with the Stanford Three Strikes Project and the Office of the State Public Defender‘s Indigent Defense Improvement Division to expand access to the tool and help practitioners identify possible resentencing candidates. It also plans to conduct a longitudinal study to identify patterns in sentencing by offense category and type to help pinpoint racial and gender biases. Lastly, Redo.io will leverage machine learning to help attorneys prioritize cases based on their past preferences and find cases similar to those they have successfully resentenced.
