المجلة الجزائرية للحقوق والعلوم السياسية
Volume 11, Numéro 1, Pages 1369-1381
2026-06-01
Predictive Criminal Justice: Expanded Legal Analysis Of Ai Systems Used To Predict Crime And Criminal Behavior
Authors : Belouadah Tayeb .
Abstract
Predictive criminal justice describes the use of artificial intelligence and statistical algorithms to forecast crime, assess individual “risk,” and guide decisions in policing, pretrial release, sentencing, corrections, and supervision. While advanced analytics may enhance consistency and resource allocation, these systems raise acute rule-of-law concerns because their outputs can function as grounds for coercive state action. Training data frequently reflect unequal enforcement patterns and underreporting, enabling feedback loops that intensify surveillance in already over-policed communities and reproduce discriminatory effects through proxy variables even where protected traits are excluded. Model opacity, trade-secret claims, and technical complexity may impair disclosure, adversarial testing, and judicial oversight, thereby undermining due process and the right to a meaningful challenge. The principal legal issue is whether probabilistic forecasts can be reconciled with constitutional standards of individualized suspicion, the presumption of innocence, transparency, accountability, and non-discrimination, or whether legality requires stringent statutory limits, auditing, and effective remedies.
Keywords
Predictive ; Criminal ; Justice ; AI Systems
Les articles similaires
Ait Ali Zaina
.
pages 584-596.
Ben Kherara Fatiha
.
Azzi Imane
.
pages 740-759.
محديد حميد
.
عنان امال
.
ص 10-25.
مزاي ليدية
.
جامعي سهيلة
.
ص 306-331.
Derrar Abdelhadi
.
Derrar Nassima
.
pages 102-110.
