This paper explores AI’s dual role in mitigating or exacerbating bias in criminal sentencing. As AI-driven tools increasingly integrate into global criminal justice systems, algorithmic justice has become a critical concern for legal practitioners, policymakers, and technologists. The study examines how AI can reduce bias through data-driven objectivity, enhanced consistency in decision-making, and comprehensive factor analysis that addresses limitations in human cognitive processing. Simultaneously, it investigates how flawed training data, problematic algorithm design, and interpretability gaps can amplify existing inequalities within judicial systems. Through comparative case studies of successful and problematic AI implementations, the research identifies key factors influencing AI’s impact on sentencing equity. It proposes practical strategies including rigorous data preprocessing protocols, systematic algorithmic auditing, and the establishment of robust legal and ethical frameworks. Findings reveal that AI’s impact on criminal sentencing is neither inherently beneficial nor harmful but is shaped by data quality, algorithm fairness, and governance structures, offering valuable guidance for responsible AI implementation in criminal justice.
Yiqiang Gao (Wed,) studied this question.