Artificial Intelligence Acceptance Among Nurses and Other Healthcare Professionals in Spain and Henan Province, China: An Exploratory Cross-Sectional Comparison

Background/Objectives: Assessing artificial intelligence (AI) acceptance may help identify questions for nursing education, but similar overall scores need not reflect similar response patterns. We compared general AI acceptance in healthcare professional samples from Spain and Henan Province, China, including a substantial nursing subgroup. Methods: This exploratory cross-sectional comparison used data from two convenience-sampled online surveys involving 7187 respondents (Spain, 3764; Henan, 3423), including 3590 nurses. We examined a five-item acceptance composite using ordinal confirmatory factor analysis, measurement invariance testing, observed and adjusted comparisons, and equivalence testing with bounds of ±0.20 pooled standard deviations. Results: Observed means were 3.59 in Spain and 3.56 in Henan; the difference of 0.028 points fell within the exploratory equivalence bounds (90% CI −0.010 to 0.067; bounds ±0.195), whose practical significance has not been established. Metric and scalar constraints met change-in-fit criteria, although absolute fit was mixed and strict invariance was not supported. The scalar model yielded an exploratory latent Henan–Spain contrast of 0.227 reference-group standard deviations (95% CI 0.152 to 0.302), which requires caution given the mixed model fit. The adjusted composite difference was 0.114 points (95% CI 0.054 to 0.173). Item-level differences ran in opposing directions. In the exploratory nursing subgroup analysis, the adjusted Henan–Spain difference was 0.081 points (95% CI −0.014 to 0.176; Holm-adjusted p = 0.283). The interval included zero and positive differences; nurse-specific equivalence was not tested. Conclusions: Similar observed averages concealed heterogeneous responses in these samples. The findings support examining distinct aspects of acceptance when designing nursing-focused needs assessments, not assuming competence or readiness for a particular AI tool. Non-probability recruitment and incomplete cross-cultural validation limit generalization and interpretation. Registration: The study was not registered.