Trust, Risk, and Uncertainty Surrounding the Use of AI in Health Care: Scoping Review of Studies on Public Perception

Background While AI has existed for decades, the introduction of generative AI tools and subsequent advancements in their capability and public use resulted in a shift in public awareness and perceptions of AI. In response, the research community has grown increasingly interested in understanding public perceptions of AI use in various contexts, including health care. Public perceptions shape AI acceptance and consequently integration into health care. It is thus crucial to understand public trust in AI and perceptions of risk as we consider its implementation moving forward. Objective This study aimed to (1) document the extent of what is known within the literature regarding public perceptions of AI applications in health; (2) interrogate existing definitions and theories of trust being used to guide investigations of public trust in AI; and (3) identify practice and policy considerations and support the design, deployment, and regulation of these technologies in a manner that is trustworthy. Methods The Arksey and O’Malley scoping review framework and the PRISMA-ScR (Preferred Reporting Items for Systematic Reviews and Meta-Analyses extension for Scoping Reviews) guidelines were used to review publications between 2022 and 2024 across 4 databases, resulting in the inclusion of 41 papers for review. Key findings were then charted and organized into tables to help answer the proposed research questions. Results Public trust in AI in health care was shaped by discourse relating to AI performance and concerns related to depersonalization, loss of human connection, privacy, bias, and transparency. Trust was found to be higher when AI was supporting rather than replacing health care professionals and was influenced by emotional responses, anthropomorphism, demographic factors, and institutional trust. Furthermore, trust extended beyond algorithms to software interfaces, providers, developers, and governing institutions, highlighting the importance of governance and oversight. Conceptualizations of trust varied widely, and trust was defined through vulnerability, competence, benevolence, and integrity. Different theories identified suggested that trust operates across individual, social, institutional, and technological levels. Finally, trust was closely linked to risk and uncertainty, where perceived risk reduced acceptance, while AI literacy, transparency, and explainability mitigated concerns and fostered trust. Conclusions Trust in AI is not uniform, and as such, efforts to build and measure trust must account for these nuances to effectively mitigate risk and support informed, ethical AI integration and deployment. On the basis of the findings of this review, we propose a multilevel conceptual framework that provides a general understanding of public trust in AI within a health care context that may guide research and policy and practice considerations. Our findings highlight the importance of considering public trust when implementing AI in health care and may support AI communication strategies, policy development, and trustworthy AI development.