The rapid advancement of artificial intelligence (AI) technologies, particularly the proliferation of intelligent systems across cyberspace and digital platforms, has enabled these technologies to play not only beneficial roles but also significant roles in facilitating and intensifying social unrest. This development has generated new challenges regarding the attribution of criminal liability. The present study aims to assess the feasibility of imposing criminal liability on artificial intelligence systems in the context of social unrest and to examine the legal foundations for attributing such liability. Employing a descriptive-analytical approach based on library research, the study addresses the following question: despite the absence of legal personality and independent will on the part of AI systems, can criminal liability arising from their interventions in situations of social unrest be justified under existing principles of criminal law and attributed to the actors involved throughout their life cycle? The findings indicate that the role of AI systems in social unrest can be analyzed at three levels—causation, facilitation, and escalation—while their modes of intervention may be examined from both structural and behavioral perspectives. Furthermore, criminal liability in this context is directed less toward AI systems themselves than toward the developers, operators, and beneficiaries who, in light of the degree of control exercised, the foreseeability of risks, the nature of their intervention, and the causal relationship between their conduct and the criminal outcome, play a substantial role in the occurrence or escalation of criminal behavior. Accordingly, the study emphasizes the necessity of reinterpreting certain traditional concepts of criminal liability and moving toward the development of legislative frameworks tailored to the distinctive characteristics of AI systems.