GoKawiilMichigan Prisons to Pilot Recidiviz AI Tool for Inmate Risk Classification
The Michigan Department of Corrections is launching a pilot program in December using an AI system from nonprofit Recidiviz to help classify incarcerated people into one of five security housing levels. The tool, which MDOC says could improve prediction of future violence risk by about 30 percent, will supplement decisions previously made by facility staff committees appointed by wardens.
GoKawiil's interpretation of the reporting above, not reported fact.
Housing classification determines which programs, resources, and populations an incarcerated person is exposed to, so shifting this decision toward an algorithm could significantly affect people's daily conditions and rehabilitation prospects. MDOC's 30 percent improvement figure is not explained, which raises questions about how the claim was validated and whether the system could introduce new biases or errors into high-stakes decisions.
- Michigan is piloting Recidiviz's AI software in multiple men's prisons starting December to classify inmates by risk level.
- The tool is meant to supplement or replace staff-led security classification committees previously run by wardens.
- MDOC claims a 30 percent improvement in predicting future violence risk, but has not disclosed how that figure was calculated.
Source: futurism.com — Joe Wilkins, 2026-11-01
Published there as: “Prisons Using AI to Predict When Incarcerated People Will Commit Future Crimes”
Read the original report → The summary and analysis above are GoKawiil's own, written from reporting by the source above. Facts and quotes belong to the original publisher.