Damages Liability for Harm Caused by Artificial Intelligence – EU Law in Flux

Abstract

Artificial intelligence (AI) is an integral part of our everyday lives, able to perform a multitude of tasks with little to no human intervention. Many legal issues related to this phenomenon have not been comprehensively resolved yet. In that context, the question arises whether the existing legal rules on damages liability are sufficient for resolving cases involving AI. The EU institutions have started evaluating if and to what extent new legislation regarding AI is needed, envisioning a European approach to avoid fragmentation of the Single Market. This article critically analyses the most relevant preparatory documents and proposals with regard to civil liability for AI issued by EU legislators. In addition, we discuss the adequacy of existing legal doctrines on private liability in terms of resolving cases where AI is involved. While existing national laws on damages liability can be applied to AI-related harm, the risk exists that case outcomes are unpredictable and divergent, or, in some instances, unjust. The envisioned level playing field throughout the Single Market justifies harmonisation of many aspects of damages liability for AI-related harm. In the process, particular AI characteristics should be carefully considered in terms of questions such as causation and burden of proof.

Keywords: Artificial intelligence (AI), EU law, Harmonisation, Damages liability, Consumer protection