Prison data analysis finds choosing lowest reoffending risk scores preserves overall accuracy
Criminal JusticePublic PolicyData Science
THE AI ANGLE
Predicting prisoner reoffending riskResearchers analyzed over 17,500 prison releases in Catalonia to study machine learning models that assess recidivism risk. They found that selecting the lowest score among equally accurate models does not reduce overall predictive performance. This gives criminal justice practitioners and policy analysts a concrete rule for handling conflicting risk assessments.
THE TEACHING ANGLE
Equally accurate models can assign different risk scores to the same prisoner, which forces policymakers to decide how to handle model disagreement.Read the original at phys.org Generate teaching or study materials
More in Criminal Justice
- What AI regulation could learn from the most hazardous industries in the worldThe Conversation — AI · November 18, 2026
- A third of the global AI supply chain is 'invisible' to governments, study findsPhys.org — Technology · November 18, 2026
- Voters Reject ALPRs and Surveillance Tech at the Ballot BoxEFF Deeplinks · November 18, 2026
- OpenAI sued by California after admitting to 80 loss-of-control incidentsArs Technica · November 18, 2026
- California DMV bars Waymo AVs from operating on the state’s freewaysThe Robot Report · November 18, 2026