Artificial intelligence, as a new and evolving technology, has been troubling lawyers in all fields in recent years. It is no different when it comes to assessing how the law will react to the harm caused by an artificial intelligence system. The present thesis first outlines the main characteristics of artificial intelligence systems that distinguish this technology from various "simple" machines and devices, i.e. its autonomy, unpredictability and the interconnectedness. Through the prism of existing non-contractual liability regimes, with the focus on the operators of such systems as the ones responsible for the damage - the subjects with most control over and benefits from the AI system, the thesis further explores the possibilities of a legal response to the damage arising from such systems. By analysing the possibilities offered by the well-known types of civil liability in Slovenian law, as well as more broadly in comparative law, the thesis attempts to find possible solutions to fill the existing liability gaps that artificial intelligence is unveiling. On the other hand, it comments on the solutions already proposed, both those put forward by the European institutions as well as those emerging among the leading legal theorists in the field.
|
