智慧型服務信任與知覺價值模型建置與驗證：以復健醫療為例

From a psychological standpoint, the level of trust that user has for AI services (e.g. AI-based customer service) depends on the levels of credibility, benevolence, and risk experienced by the user during the service-rendering process. The current inference mechanism of AI services is not transparent, causing users to have low levels of trust in AI services. This is especially evidentamong high age groups who did not grow up alongside artificial intelligence. Research has shown that increasing a user’s trust in AI services requires comprehensible technological and decision-making frameworks that can be explained and verified. For this reason, this study proposes a theoretical model for the integration of human-computer interaction and scientific literacy toincrease user trust in AI. The principal components of this model comprises the disclosure of information on the AI’s decision-making process, the delineation of its inference process, and the incorporation of user feedback to help users feel more included in the process. The methodologies that are used include data visualization analyses as well as the establishment of user feedback and verification mechanisms. This study first looks at AI rehabilitation services developed by the partner hospital followed by data visualization analyses which are used to guide users in understanding the AI’s inference process. Then, therationality of the service output data is evaluated. Finally, interviews are conducted with users to evaluate and give feedback on the experiences with physical fitness examination AI services as a way to obtain a holistic understanding of AI service efficacy. The results of this study contribute to the understanding of the factors of human-computer interaction as well as the effectof AI literacy in increasing user trust in AI services.