Odgovornost za štetu prouzrokovanu upotrebom sistema veštačke inteligencije u pravu Evropske unije – stanje i izazovi

The purpose of this article is to explore the regulation of AI and the associated liability within the European Union (EU). It focuses on the legislative approach of the EU, particularly examining the AI Liability Directive. The discussion begins by outlining the necessity for regulation and liability in the context of AI systems. Further, the author analyses the regulatory purpose and the legislative approach. In the second part, AI is analysed through a riskbased approach, considering both ethical and legal standards essential for ensuring the trustworthiness of AI. The characteristics of AI and the levels of risk associated with its deployment are also explored. The compliance requirements established by the AI Act of the EU are scrutinized, offering insights into the standards set for AI systems to be deemed reliable and safe. This is followed by a review of the issues surrounding liability for damage caused by AI systems. Key aspects include identifying the parties involved in potential cases—claimant and defendant—the nature of the damage, the causal link, and the targeted rebuttable presumption of causality as well as the presumption of non-compliance with the conformity requirements as a reflection of the fault-based liability system. The article concludes by discussing the perceived ‘open issues’ in creating a comprehensive liability regime for AI within the EU and their effect on the national liability systems. It addresses the challenges of approximating national liability laws with EU law, as well as the implementation and enforcement of the rules.