Victim Compensation Laws and Restorative Justice

Abstract

Victim compensation has become an increasingly important component of contemporary criminal justice, yet monetary relief and restorative justice frequently operate as institutionally separate processes.  This study examines whether explainable artificial intelligence can support a more consistent, timely, and victim-centred assessment of compensation and restorative needs without replacing judicial or administrative discretion. The identified research gap concerns the limited integration of statutory compensation criteria, multidimensional victim harm, rehabilitation requirements, and restorative outcomes within computational decision-support models.  An Explainable Victim Redress Assessment Framework (EVRAF) is proposed to classify compensation-priority levels and estimate broader restorative-support requirements from structured and textual case information. The framework conceptualizes physical injury, psychological impact, economic loss, dependency, treatment requirements, livelihood disruption, vulnerability, and rehabilitation needs as interconnected dimensions of harm. A hybrid architecture combining interpretable machinelearning classification with natural-language processing and explanation mechanisms is proposed to preserve human oversight and make recommendations auditable. Rather than predicting what a victim legally “deserves,” the system is designed to identify potentially comparable cases, flag inconsistencies, and provide structured evidence for review by authorized decision-makers.  The research contributes an interdisciplinary model connecting victimology, restorative justice, explainable AI, and digital legal governance while emphasizing fairness, privacy, transparency, and meaningful human review.