Criminal Liability for the Acts of Autonomous Self-Learning Robots in the Absence of an Express Criminalization Provision

Abstract

The world is witnessing rapid developments in the field of autonomous self-learning robots, giving rise to a profound legal issue concerning criminal liability for harmful acts committed by such robots in the absence of an express criminalization provision. This study aims to analyze this issue, identify manifestations of legislative gaps in comparative legal systems, evaluate the various theoretical approaches to attributing liability, and propose a synthesized framework for addressing it. The study adopts a comparative analytical approach, complemented by inductive and deductive methods. The study reached several main conclusions. First, existing criminal provisions are insufficient to address acts committed by autonomous robots because they were designed to regulate human conduct or traditional automated behavior. Second, recognizing an electronic legal personality for robots remains the subject of profound theoretical disagreement and has not been translated into binding legislative provisions. Third, the collective liability of a legal person is not a comprehensive solution, as it implicitly presupposes the existence of attributable human fault. Fourth, the optimal solution lies in a hybrid model that combines specific legislation, broader judicial interpretation, and flexible liability-allocation mechanisms. Fifth, criminal liability in this field should be understood as an evolving framework for distributing responsibility among multiple parties, rather than as a rigid concept attributed to a single actor.