Razpletanje vozla civilne odgovornosti v zdravstvu ob uporabi umetne inteligence

slovenian

2026

volume 81, issue 3/4

artificial intelligence   burden of proof   duty to explain   healthcare   joint and several liability   liability for damages   medical standards   strict liability   therapeutic freedom

Univerza v Ljubljani, Pravna fakulteta
(Obvezni izvod spletne publikacije)

The article examines civil liability in healthcare arising from the use of robotics and artificial intelligence (AI) and shows how digitalisation is reshaping traditional questions of medical negligence, the duty to inform, and the burden of proof. Its central thesis is that, in principle, Slovenian law reveals no specific liability gap inherent in AI: the core doctrines of contractual and tortious liability, employer liability for the acts of healthcare staff, and product liability remain applicable even in digitised forms of treatment. AI nevertheless significantly intensifies practical difficulties, particularly in proving causation and allocating liability across the chain of actors involved (the doctor, the healthcare institution, the manufacturer, the developer, and the provider responsible for maintenance and updates), while the opacity of algorithmic decision-making further complicates matters (the “black box” effect). The article first highlights the structural weaknesses of the Slovenian healthcare system, particularly the inadequate recording of safety incidents and the lack of uniform clinical guidelines, and considers how these deficiencies are reflected in compensation litigation. It then analyses the legal permissibility of the use of AI through the lens of therapeutic freedom and medical standards, emphasising that the duty to inform is particularly demanding when new methods are used. Finally, the article considers possible de lege ferenda solutions, in particular whether the application of strict liability and joint and several liability is justified in cases where, owing to technological and organisational complexity, the liable party cannot be identified with sufficient certainty

20.07.2026