The mediating role of job crafting in the relationship between perceived organizational support and nurses' attitudes towards artificial intelligence: a cross-sectional study based on the job demands–resources model

Background Artificial intelligence (AI) is being rapidly integrated into clinical nursing, yet many nurses remain ambivalent or resistant to these technologies. While perceived organizational support (POS) and job crafting have each been linked to nurses’ work attitudes, little is known about how they jointly shape nurses' views on AI. Drawing on the job demands–resources (JD-R) model, in which perceived organizational support represents a job resource and job crafting a proactive behaviour, this study examined whether job crafting mediates the association between POS and nurses' attitudes towards AI application. Methods Between April and June 2026, 314 clinical nurses from a tertiary grade A hospital in China were invited via convenience sampling; 305 returned valid responses. Participants completed the Perceived Organizational Support Scale (POSS), the Nurse Job Crafting Scale (NJCS), and the Chinese version of the Attitudes towards AI Application at Work (AAAW) scale; items 14–21 of the AAAW were reverse-scored. Mediation was tested using the SPSS PROCESS macro (Model 4) with 5,000 bootstrap samples. Results Mean item scores were 3.86 (SD = 0.80) for POS, 3.90 (SD = 0.59) for job crafting, and 3.46 (SD = 0.68) for AI attitudes. All three variables showed significant positive correlations (r = 0.502–0.715, all p < 0.01). Job crafting partially mediated the POS–AI attitudes relationship, accounting for 60.8% of the total effect (indirect effect = 0.248, 95% CI [0.169, 0.337]). The direct effect of POS remained significant after controlling for the mediator (b = 0.160, p = 0.006). Because AAAW is multidimensional, the model was also tested on theoretically distinct facets; the indirect effect was significant for positive perceptions (0.301, 95% CI [0.203, 0.403]) but not for AI-related concerns. Conclusions Job crafting was associated with the relationship between POS and nurses’ AI attitudes in a pattern consistent with partial mediation. Organizational support and job crafting were associated with more positive attitudes toward AI, consistent with the JD-R model. Both factors may matter for AI adoption in nursing. The cross-sectional design limits causal inference, and longitudinal studies are needed to confirm these pathways.