CIVIL LIABILITY FOR ARTIFICIAL INTELLIGENCE IN MEDICINE: IS THERE A NEED FOR A NEW PARADIGM?

Main Article Content

Petr Šustek
Martin Šolc

Abstract

The rapid integration of artificial intelligence (AI) into medical practice raises pressing questions for civil liability law. While AI systems promise enhanced diagnostic and therapeutic capabilities, their autonomous and opaque black box nature pose unique legal challenges. This paper investigates whether existing tort law frameworks, particularly within the Czech legal system, are sufficient to address harms caused by medical AI. It systematically analyses three core liability regimes: breach of the standard of care, strict liability based on a defect of a thing or product, and liability for damage caused by a thing in and of itself. The authors argue that, although each of these regimes presents interpretive and evidentiary difficulties in the context of AI, none is fundamentally unfit for addressing AI-related harms. Instead of advocating for a wholly new legal regime or the creation of a fictive electronic person, the paper supports a cautious evolution of existing doctrines, refined through case law, medical guidelines, and professional standards. Special attention is given to the distinction between certified medical AI devices and general-purpose systems like large language models. The authors also explore the implications of EU legislation, notably the AI Act. The paper concludes that legal conservatism, rooted in pragmatism, may serve the justice system best as the practical use of AI in medicine continues to unfold.

Article Details

Section
Articles