Between the Algorithm’s Recommendation and the Individual’s Will: Two Practical Case Studies on AI in Healthcare and Elder Care

Abstract

This article reworks the discussion on artificial intelligence in healthcare and elder care from a predominantly conceptual debate into a case-oriented legal and bioethical analysis. Its central claim is that the most important practical risks do not arise where an AI system formally replaces the human decision-maker, but where it quietly restructures the conditions under which patients, professionals, relatives and care institutions act. The article therefore analyses two model case studies grounded in real and recurring patterns of practice: first, the refusal of life-sustaining treatment in the presence of an algorithmic recommendation for escalation of therapy; second, the gradual intensification of AI-supported home monitoring of an older person in the name of safety. In both cases, the decisive issue is not whether the technology is technically accurate, but whether legal and administrative arrangements preserve meaningful autonomy, contestability, proportionality and responsibility. The argument proceeds by connecting practical case analysis with medical-law principles concerning competent refusal, informed consent, dignity, privacy and the continuing relevance of the dignity of risk. The article concludes that a rights-sensitive use of AI requires procedural safeguards embedded in institutional design: functional transparency, documented dialogue, renewable consent, protected deviation from automated outputs, and accessible routes of contestation.