A new open-access article entitled “Hospitals as AI Deployers: Organisational Failures and Direct Liability Risks”, authored by Claudia Giorgetti, Giuseppe Contissa and Arianna Giorgetti, has been published in the European Journal of Risk Regulation.
The article examines the direct liability risks faced by hospitals when deploying AI-based medical devices, particularly clinical decision support systems (CDSSs). While discussions on AI liability in healthcare traditionally focus on manufacturers and physicians, this research shifts attention to hospitals and their potential exposure to fault-based liability arising from organisational failures.
Drawing on the EU AI Act, the Medical Device Regulation and Health Technology Assessment literature, the study identifies five phases in the AI deployment lifecycle and maps fourteen organisational failures that may occur throughout these phases, including inadequate due diligence, ineffective oversight, deficiencies in staff training and poorly designed clinical workflows.
The article further explores the shared responsibilities of hospitals and AI providers in shaping the clinical environment, highlighting how organisational decisions can influence the effective and safe use of AI systems by healthcare professionals.
Finally, the study assesses the circumstances in which organisational failures may give rise to direct liability, particularly where hospitals breach applicable AI Act deployer obligations, repeat organisational shortcomings or fail to address previously identified risks.
The article is available open access:
Giorgetti, C., Contissa, G., & Giorgetti, A. (2026). Hospitals as AI Deployers: Organisational Failures and Direct Liability Risks. European Journal of Risk Regulation, 1–17.
Read the full article: https://doi.org/10.1017/err.2026.10176