Il giudizio etico oltre il protocollo: intelligenza artificiale, responsabilità e trasformazione biogiuridica dei Comitati etici nella ricerca sanitaria. (Una lettura del rapporto OMS 2026 tra filosofia del diritto e governo dell’innovazione)

The 2026 World Health Organization report “Artificial intelligence-related health research: ethics review and oversight” is used here as the point of departure for a broader biolaw thesis: AI destabilizes not only the object of ethics review, but also the temporal and institutional conditions under which normative responsibility is attributed. The article distinguishes technical reliability, legal compliance and ethical justification, arguing that none can substitute for the others. It develops a model of dynamic ethics review based on five material-change triggers concerning data, model, population, purpose and impact. It further proposes a threefold principle of non-delegability of ethical judgment: institutional attribution of the decision, ownership of the reasons supporting it, and effective power to revise it when circumstances change. Against this background, the research ethics committee is reconceived as an institution of normative mediation and responsible human oversight. The European framework—GDPR, EHDS, AI Act and the Council of Europe Framework Convention on AI—shows why ethical relevance cannot be reduced to the scope of positive regulation, especially where scientific research falls outside or precedes specific AI Act obligations.Keywords: Artificial intelligence; Research ethics committees; Biolaw; Dynamic ethics review; Non-delegability of ethical judgment.