Towards a European AI liability system

Abstract

This essay deals with certain civil liability implications of artificial intelligent systems in the light of the recent steps taken by the European Union. In order to create not only an ethical but also a lawful AI, the EU strives to lay down the framework of the future common liability rules for damages and harms caused by any application of AI technology. The Commission’s new Proposal (Artificial Intelligence Act, AIA) reflects on an innovative approach to the regulation which can tackle with the special features of the AI systems, lays down rules according to the risk management approach and the class-of-application-by-class-of-application approach. In this essay, the strict-based liability for high-risk AI systems and the concept of the frontend and backend operators are in the focal point.