Researchers suggest using lowest risk scores for prisoners in Catalonia
Ashwin Singh and Carlos Castillo presented findings at a conference in Malmö regarding the RisCanvi system used in Catalan prisons since 2009.
Analysis of over seventeen thousand releases
The research team examined the RisCanvi system which has been operational in Catalan prisons for more than fifteen years. They analyzed a dataset comprising more than 17,500 prison releases to determine recidivism outcomes after release. The study compared several equally reliable computational models that produced different risk ratings for the same individuals.
Policy favors the person affected
Ashwin Singh from TU Wien and Carlos Castillo from ICREA suggest using the lowest value when multiple models agree. This approach resolves ambiguity in favor of the individual concerned rather than selecting a higher score by chance. The researchers translated complex legal rules into an algorithm to systematically determine whether a person had reoffended legally.
Comparisons include gender and nationality
Analyses contrasted males and females alongside those who held or lacked Spanish citizenship. Technical actions guaranteed that rehabilitation advancements achieved while serving time reduce the risk rating as intended. The final decision on parole remains with human professionals who use these probability assessments to support their choices.
Reported by one outlet
Only one outlet has published this. Nothing here has been checked against a second report, so read it as that outlet's account and follow the link below for the original.
Reported by
1 independent outlet. Headline as published. Links open the original report.