Knowledge and attitudes regarding AI-assisted documentation among clinical nurses in China: a cross-sectional study

BACKGROUND: China’s ageing population, shortage of nursing staff and rising care demands have created a substantial documentation burden for clinical nurses, undermining care quality and nurses’ physical and mental health. AI is widely viewed as a potential means to reduce this burden, yet nurses’ knowledge of and attitudes toward AI-assisted documentation remain poorly characterised. OBJECTIVE: To characterise Chinese clinical nurses’ knowledge and attitudes regarding AI-assisted documentation and to assess associations with gender, age, education and AI-related training. METHODS: We conducted a cross-sectional online survey of clinical nurses in 15 tertiary hospitals across seven provincial-level administrative regions in China using convenience sampling. A self-administered questionnaire, developed from a literature review and Delphi expert consultation, comprised three sections-demographic information, knowledge of AI-assisted documentation and attitudes-with 21 items in total. Data were analysed using descriptive statistics, t tests or one-way analysis of variance and multiple linear regression. RESULTS: Of 600 returned questionnaires, 492 were valid (effective response rate 82.0%). The mean AI knowledge score was 10.74 ± 5.04 (maximum 20), indicating generally low knowledge. Nurses showed relatively high awareness of risks, such as fabricated AI-generated content and the need for professional review of AI outputs, but limited understanding of prompt design and basic tool functions. The mean attitude score was 24.48 ± 10.00 (maximum 44), indicating an overall cautious and conditional acceptance attitude: most nurses were willing to use AI for documentation in the future but expressed marked concerns about privacy protection and the originality of AI-assisted content. Univariate and multivariable regression analyses identified age, education and AI-related training as significant correlates of knowledge and attitudes (p < 0.05), with training showing the strongest association (p < 0.001). Knowledge and attitudes were positively correlated (p < 0.001). CONCLUSION: Chinese clinical nurses had limited knowledge and elevated risk perception regarding AI-assisted documentation but reported cautious and conditional acceptance attitudes overall. AI-related training was associated with higher knowledge and more positive attitudes; however, given the cross-sectional design, this association may reflect prior interest or exposure rather than a causal effect. CLINICAL TRIAL NUMBER: Not applicable.