The integration of artificial intelligence (AI) into criminal sentencing has increasingly influenced decision-making within contemporary criminal justice systems. In Malaysia, the implementation of the Artificial Intelligence Sentencing System (AISS) in Sabah and Sarawak courts has intensified debates concerning the compatibility of AI-assisted sentencing with the rule of law, judicial discretion and constitutional safeguards. Although AI-powered sentencing tools are promoted as mechanisms capable of improving efficiency, consistency and predictive assessment, their application raises serious concerns relating to transparency, explainability, procedural fairness and algorithmic bias. The absence of a specific Malaysian legal framework regulating AI-assisted sentencing further creates uncertainty regarding the extent to which courts may rely on algorithmic recommendations without compromising the accused person’s constitutional rights under Articles 5 and 8 of the Federal Constitution. Accordingly, this study examines the legal implications of AI-assisted criminal sentencing in Malaysia through a doctrinal legal research methodology. The research analyses relevant statutory provisions, including the Penal Code (Act 574), Criminal Procedure Code (Act 593), Subordinate Courts Act 1948. The study finds that excessive reliance on opaque algorithmic systems may undermine natural justice, weaken judicial reasoning and challenge the principles underpinning the rule of law. This article argues that AI should operate solely as a supplementary judicial aid subject to strict human oversight, transparency requirements and explainability safeguards within Malaysia’s criminal justice framework.