KJCKorea
Journal Central

To apply healthcare AI ethics in clinical and research settings, self-regulatory frameworks emphasize the necessity of external validation. Building on theoretical discussions that justify patients and the public as key actors in this validation process, this article conducts a scoping review on patient and public involvement (PPI) in healthcare AI and proposes a model for their engagement. A search was conducted in academic databases using the keywords “AI” and “patient and public involvement.” Out of the 114 studies that were identified, 16 were selected for review; the selected studies consisted of four scoping reviews, eight empirical studies, and four theoretical papers. The review confirmed ongoing discussions on PPI in healthcare AI, but highlighted the lack of detailed implementation strategies. This article suggests participatory design, healthcare AI literacy, and citizen science models as potential approaches for engagement, discussing their possibilities and limitations.

With the rapid advance of artificial intelligence (AI) technology and the development of healthcare AI, there is a growing need for reflection and debate on the meaning, scope, and feasibility of participation in healthcare AI. This review contributes to this process of reflection in the following three ways. First, it explores efforts to define AI as intelligent technology. Second, it explains the concept of participatory AI, which applies participatory design principles to the development of AI systems. Finally, it reviews critical concerns with healthcare AI, including the risks of “participationwashing” and the need for a context-sensitive understanding of participation in AI systems.

Amid the growing prominence of healthcare artificial intelligence (AI), current research still pays only nominal attention to patient and citizen participation. However, this engagement rarely extends beyond tokenism, as healthcare AI ethics continues to reinforce rigid distinctions between doctors and patients and also between subjects and objects. These distinctions, which impedes genuine participation, are rooted in a specific ontology. By embracing a new ontology, participation becomes self-evident, shifting the focus from questioning its feasibility to justifying any limitations on its scope. Importantly, this shift is not limited to healthcare AI ethics; adopting a new ontology could enable meaningful progress across the entire field of bioethics.