Предлог Директиве о одговорности за вештачку интелигенцију
The proposal for a Directive on liability for artificial intelligence

Summary/Abstract: Although there are a large number of initiatives, both at the level of national jurisdictions and the global level (proposals on international agreements or the establishment of an international monitoring body), so far, only the European Union (EU) has made any serious attempt to protect citizens from the risks associated with the use of artificial intelligence systems. The European Act on Artificial Intelligence entered into force on August 1, 2024. In the discussions about its content, one of the main criticisms was directed at the lack of individual remedies. In the meantime, the European Commission drafted the Directive on Liability for Artificial Intelligence, whose provisions are the subject of this paper. The proposal of the Directive is considered in its entirety in order to indicate the importance of the matter in question and provide a basis for further research. It follows from the analysis that if the current text of the Proposal for the Directive is not amended, the injured party could face obstacles in proving the existence of a cause-and-effect relationship in the damage caused, especially when it comes to artificial intelligence systems that are not classified as high-risk by the Act on Artificial Intelligence. The rules on non-contractual civil liability in connection with artificial intelligence remain in the positions of subjective liability (responsibility based on fault) to facilitate the position of injured persons. Such an approach was also confirmed through the phased implementation process of the Proposal for the Directive because it was only in the second phase that the possibility to move from subjective to objective responsibility for damage was projected.

• Details
• Contents