AbstractsLaw & Legal Studies

The present thesis explores questions related to the state responsibility for autonomous systems. The aim of this thesis is to conceptualise and elaborate on the main state responsibility issues related to the autonomous systems. The analysis consists making a definition of autonomous systems, descriptive examination of the state responsibility doctrine in the present context and an attempt to concretise the phenomenon in the scenario analysis. For the purpose of the thesis, autonomous systems represent technical systems that have somewhat extensive independent decision-making capabilities. There are different categories of these systems and so called off-the-loop systems are the purest example of them. In addition to independence in decision-making processes, the capability to learn and to adapt in some extent in unpredictable circumstances is also typical for autonomous systems. All these features are assumed to be constructed by humans and thus at the meta level they cannot be fully separated from human influence. It is also relevant to understand that autonomous systems contain a group of technological apparatuses which through inter-communication form a coherent system. The defining and important matter from the legal perspective is the nexus between these systems that are gaining autonomy and the responsibility for those actions. It is the view taken in the thesis that contemporary state responsibility system is best described in the International Law Commission’s draft articles on the Responsibility of States for Internationally Wrongful Acts. Following this premise, the research is concentrated on the topics arising from these articles. The main areas of the analysis are general applicability, attribution, state agent, circumstances precluding wrongfulness and ultra vires. Moreover, the framework included rundown of the due diligence and liability aspects of autonomous systems. Pursuant the thesis, it can be concluded that state responsibility includes an international wrongful act and attribution of that act to a state. In the autonomous systems context there is a huge number of primary norms and any breach or omission will lead to an international wrongful act. Attribution to a state is always based on a human or a group of humans. The autonomous systems cannot create the legal link of attribution as such. The circumstances to preclude wrongfulness can be applied to the autonomous systems as well. Additionally, actions committed by autonomous systems that include components from several states can lead to shared responsibility. It is also by definition that autonomous systems cannot act ultra vires in legal sense. Furthermore, in cases when the action of a state is not against an international norm but still causes damage, the liability regime might be applicable – also in the context of autonomous systems. The real life analysis of autonomous systems is always context specific and the chapter on scenario analysis will provide some examples of the state responsibility questions of the different stages…