Artificial Intelligence in Sentencing Decisions: Legal, Ethical, and Human Rights Considerations

Abstract

The use of artificial intelligence (AI) in sentencing decisions is increasingly discussed as part of modern criminal justice reform. In Indonesia and other jurisdictions, AI-based tools are proposed to assist judges in predicting recidivism, assessing risk, and improving sentencing consistency. However, such developments raise significant legal, ethical, and human rights concerns. This article examines the implications of AI-assisted sentencing within the Indonesian legal system, focusing on compatibility with principles of judicial independence, fairness, and due process. Using normative legal research and comparative analysis, the study evaluates the risks of algorithmic bias, lack of transparency, and over-reliance on automated decision-making. The findings suggest that AI systems may reinforce existing social inequalities embedded in training data, potentially undermining individualized justice. Furthermore, the absence of clear regulatory standards in Indonesia increases uncertainty regarding the admissibility and legitimacy of AI-generated recommendations in sentencing. The article argues that while AI may enhance efficiency, sentencing authority must remain fundamentally human-centered. It proposes strict regulatory safeguards, including explainability requirements, judicial oversight, and ethical guidelines for AI use in courts. The study concludes that AI in sentencing should be limited to advisory functions to preserve constitutional and human rights principles.