This article explains the risks that arise from AI's ability to change the intrinsic nature of how health care is delivered. Such potential AI transformations raise ethical risks that are normative, epistemic, or overarching. Such risks relate to inconclusive, inscrutable, or misguided evidence; unfair outcomes or transformative effects; or traceability.

They point out that current responses to the ethical issues are centered on protecting the individual. However, ethical codes do not take into account the epistemic, normative, and overarching ethical risks associated with AI use in health care that arise at the relationship, group, institutional, and societal levels.

They give some examples to illustrate these concepts.

Note that this article is available free of charge and you can register for free.