<![CDATA[Contemporary criminal law faces a significant paradigmatic challenge with the emergence of Artificial Intelligence (AI) as a non-human entity capable of autonomous action and producing legally relevant consequences. The traditional criminal law system—rooted in an anthropocentric paradigm that attributes mens rea and actus reus exclusively to human actors—has become inadequate in explaining liability within the context of autonomous and adaptive algorithmic decision-making. This study aims to analyze how the criminal law system responds to actions and decisions generated by AI that result in legal consequences, and to formulate a conceptual model of AI criminal liability that ensures legal certainty, justice, and accountability. The research employs a normative legal method with a conceptual approach. The findings reveal that the criminal law system encounters both a culpability gap and a liability gap due to the absence of a legal subject that can be held directly accountable for AI’s actions. A reconstruction of the criminal law paradigm is therefore necessary through the adoption of a hybrid criminal liability framework that integrates human, corporate, and AI accountability based on the degree of control and risk creation. This study recommends limited recognition of electronic personhood for certain AI entities, alongside the application of risk-based accountability and the precautionary principle as new normative foundations for establishing an adaptive, accountable, and just criminal law system in the digital era]]>
