Network analysis of artificial intelligence attitudes and literacy among clinical nursing educators: Subgroup comparisons based on training experience and educational implications.

BACKGROUND: Artificial intelligence (AI) is being rapidly integrated into healthcare practice. However, the complex interrelationships between AI attitudes and literacy among clinical nursing educators, and the impact of training experience, remain poorly understood. OBJECTIVE: To delineate the network structure of AI attitudes and literacy among clinical nursing educators, identify core and bridge nodes, and compare networks between subgroups with versus without AI training experience. DESIGN: A cross-sectional survey using convenience sampling. SETTING: 46 healthcare institutions across 26 provinces in mainland China. PARTICIPANTS: 509 clinical nursing educators recruited from May to July 2025. METHODS: The General Attitudes Towards Artificial Intelligence Scale and the Artificial Intelligence Literacy Scale were used. A LASSO-regularised partial correlation network analysis constructed a 14-node network. Strength centrality, bridge strength, expected influence, and predictability were calculated. Network structures were compared between the group with training experience and the group without. RESULTS: The 14-node network had 59 non-zero edges (64.8% density; mean weight = 0.063). EV_3 (contextualised AI adoption) showed the highest strength centrality (1.608); ET_1 (ethical compliance) had the highest expected influence (1.359); US_3 (improved work efficiency) was the key bridge (bridge strength = 1.403). No significant group differences were found in global network structure or global strength (all P > 0.05). The training group showed higher bridge strength for EV_3, AW_2, and EV_2, while the non-training group showed higher bridge strength for ET_3 and US_2. Network stability was excellent (CS coefficient = 0.672). CONCLUSIONS: AI attitudes and literacy among clinical nursing educators form a stable, interconnected network. Contextualised application is the core driver, ethical compliance has the strongest positive activation potential, and improved work efficiency serves as the critical bridge. Training experience does not alter the global architecture but reshapes specific bridging roles. Stratified, stepwise training strategies should be prioritised for those without prior training.