The rapid advancement of artificial intelligence is bringing about profound transformations in socio-economic life while simultaneously posing significant challenges to contemporary criminal law. As AI systems increasingly acquire capabilities of self-learning, autonomous decision-making, and a high degree of operational autonomy, the determination of criminal liability in cases where artificial intelligence causes harm has become a complex and highly contested legal issue. This article examines the difficulties associated with applying traditional doctrines of criminal liability to artificial intelligence, with particular emphasis on the elements of fault, the identification of the responsible legal subject, and the causal nexus between conduct and harmful consequence. Employing doctrinal legal analysis, comparative law methodology, and case study approaches, the study evaluates existing models of criminal liability related to artificial intelligence within contemporary international legal frameworks. The findings suggest that artificial intelligence does not, at present, satisfy the criteria necessary to be recognized as a subject of criminal liability in the traditional legal sense. Instead, the article argues for the development of a hybrid liability framework, encompassing individuals as well as organizations involved in the design, development, deployment, and operation of artificial intelligence systems.