This study aimed to validate the mediating role of nurses' AI trust in the relationship between AI uncertainties and AI competence.
A cross-sectional study.
A purposive sample of 550 registered nurses with at least 1 year of clinical experience from three tertiary and two secondary hospitals in Jinan and Hangzhou, China, was used. Data were collected using structured questionnaires assessing AI uncertainty, trust and competence. Demographic data included gender, age, education level, years of clinical experience, professional title and hospital level. Mediation analysis.
Most nurses were from tertiary hospitals (88.9%), held a bachelor's degree (87.6%), and had over 6 years of experience. The mediating role of AI trust between AI uncertainties and AI competence is validated. AI uncertainties affected AI trust (B = 0.39, p < 0.0001), explaining 10% of the variance. AI uncertainties and AI trust affected AI competence (B = 0.25 and 0.67, p < 0.0001), explaining 63% of the variation. AI trust's total effect was 0.51, comprising direct and indirect effects of 0.25 and 0.26, respectively.
Hospitals can reduce uncertainty through an AI-transparent decision-making process, providing clinical examples of AI and training nurses to use AI, thereby increasing trust. Second, AI systems should be designed to consider nurses' psychological safety needs. Hospital administrators utilise optimised AI technology training and promotional techniques to mitigate nurses' resistance to AI and enhance their positive perceptions of AI competence through trust-building mechanisms.
Impact: Enhancing nurses' AI trust can reduce uncertainty and improve their competence in clinical use. Strategies such as transparency, explainability and training programmes are crucial for improving AI implementation in healthcare.
This study focused solely on clinical nurses and did not include patients or the public.
The study adhered to STROBE guidelines.