Autonomous Systems (AI) and Criminal Imputability: Challenges for Modern Law

Autonomous systems are changing the foundations of criminal liability. Where machines act autonomously, normative gaps arise that put the principle of culpability to a new test. This paper explores the question of how criminal imputability can be reconstructed under the conditions of technical autonomy. It pursues the goal of determining the prerequisites of a criminal law that can react to autonomous systems without losing its normative identity. The study follows a comparative dogmatic methodology. It combines classical legal analysis with interdisciplinary approaches to reflection and is based on contrasting case studies in the fields of transport, medicine and the military. These sectors are exemplary for high-risk areas in which technical systems directly intervene in protected legal interests and thus enable generalisable statements about responsibility and guilt. The results show that criminal imputability remains possible even in the age of artificial autonomy if it is normatively re-anchored by preventive structures. The basis remains action, causality, intent, negligence and culpability, supplemented by organisational obligations, auditability and verifiability. The AI Act and the Product Liability Directive 2024 create the framework for preventive responsibility at European level. Internationally, human rights standards, in particular Article 1 GG, Article 2 ECHR and Article 36 ZP I, ensure the legitimacy of this order. Criminal law retains its authority if it makes responsibility visible, enforces transparency and upholds human dignity as the limit of technical power to act. It remains the guarantor of human freedom in an increasingly automated world. Keywords: Autonomous systems; imputability under criminal law; artificial intelligence; culpability principle; organisational responsibility; responsibility gap; auditability; human dignity; international regulation; AI Act