Forbidden uses of AI and liability

Abstract

Adopting a risk-based approach, categorised into unacceptable risk, high risk, and minimal or no risk, along with a specific transparency risk, Article 5 of the AI Act prohibits the placing on the market, the putting into service, or the use of certain AI systems. These rules can be understood as legal provisions protecting the interests of others, which facilitates the articulation of the various prerequisites of liability, thereby opening the possibility for compensating those harmed by such practices. This carries significant practical and normative consequences, particularly regarding the possibility of advocating for the reversal of the burden of proof concerning fault and facilitating the establishment of a sphere of risk/responsibility. From this sphere, the issue of objective imputation (formerly addressed under the concept of causality) can and should be considered. In this paper, we seek to determine the scope of relevance of each of the prohibited uses provided for in Article 5 of the AI Act in order to deal with the specific issues of liability, such as wrongfulness, fault and causation.