Identifying Measurement Dimensions of Users' Benefit-Risk Perceptions of AI in Healthcare: A Scoping Review

The rapid integration of artificial intelligence (AI) into healthcare presents a double-edged nature, making systematic assessment of users' benefit-risk perceptions critical. However, a unified, multidimensional framework for such measurement is currently lacking. This review aims to systematically identify and synthesize existing measurement instruments for users' benefit-risk perceptions of AI in healthcare, and to propose an integrated framework based on the evidence. Guided by Arksey and O'Malley's 5-stage framework, we retrieved quantitative studies describing measurement dimensions for users' benefit-risk perceptions regarding AI in healthcare. The search covered 8 Chinese and English databases from their inception to December 6, 2025. Two reviewers independently performed study screening and data extraction, with subsequent synthesis and visual presentation of findings. Based on a synthesis of 49 eligible studies, we developed a measurement framework encompassing 5 benefit and 6 risk dimensions, where technological attributes often exhibit a dual nature. Current measurement instruments consistently emphasize functional benefits, cost benefits, and privacy risks across diverse healthcare contexts, user groups, and geographical regions. In contrast, social benefits and capability development risks generally receive less consideration. Furthermore, variations in instrument design are primarily reflected at the subdimension level. This framework extends classical technology acceptance theories. It provides a theoretical basis for standardized instrument development and offers guidance for the clinical implementation of AI in healthcare. Future research should explore how perceptions evolve with advancing AI maturity and clinical integration to support responsible adoption.