Abstract
The subject of the article is the issue of non-contractual civil liability, especially based on the principle of fault, related to the use of Artificial Intelligence systems. The specific characteristics of artificial intelligence systems, including their complexity, autonomy and nontransparency, may make it extremely difficult for injured parties to identify the entity bearing liability and prove compliance with the requirements for the recognition of a compensation claim. The reason for conducting scientific research in this area is the presentation by the European Commission of a draft directive on liability for Artificial Intelligence. The problem of adapting the regime of non-contractual liability in connection with the use of Artificial Intelligence systems is of extraordinary practical importance. At the same time, it is an important research issue, especially in the context of popularizing the use of Artificial Intelligence systems. The analysis sought to determine whether the directive could achieve its intended goal of ensuring effective protection for victims. This issue required examination in the light of methodological instruments classically used in the discipline of “legal sciences”, with particular emphasis on the formal-dogmatic method and an assessment of the practical dimension of the solutions proposed in the draft directive. The article is the first comprehensive discussion of the topic in Polish doctrine and is a contribution from Polish legal science to the analysis of the draft directive, as it also contains proposals for the necessary directions of changes.