Development and preliminary validation of an AI acceptance scale for intercultural teaching and its cross-sectional associations with university teachers’ occupational wellbeing

Background University teachers’ occupational wellbeing and emotional exhaustion are relevant occupational-health concerns in higher education. As generative artificial intelligence becomes embedded in teaching work, AI acceptance may represent a perceived, context-specific appraisal of whether AI is usable for intercultural teaching tasks. This study developed a preliminary measure of AI acceptance for intercultural teaching and examined its cross-sectional associations with teacher resources and occupational-health indicators. Methods This two-sample cross-sectional study used Sample A to examine the empirical behavior of the AI Acceptance for Intercultural Teaching Scale and Sample B to test measurement validity and occupational health associations. IBM SPSS Statistics 27.0.1 was used for descriptive statistics, EFA, Cronbach’s alpha, correlations, and regression analyses. CFA and SEM were conducted in R 4.6.1 (lavaan 0.6.21, lavaan.mi 0.1.0, and mice 3.19.0) using robust maximum likelihood (MLR) estimation; CR and AVE were calculated from standardized CFA loadings, and HTMT from item correlations. Hypothesis tests used non-imputed scale-score regressions; multiple imputation (mice, m = 5) with D2-pooled CFA/SEM model tests served as sensitivity analysis, and indirect-association bootstrap intervals were sensitivity summaries rather than formal D2-pooled tests. Results The AITA items were suitable for factor analysis in Sample A (KMO = 0.940), but the exploratory loading pattern did not cleanly reproduce the theorized four TAM-based dimensions. In Sample B, the overall AITA score showed good internal consistency (alpha = 0.858), whereas the four subdimensions showed substantial overlap and weaker reliability and convergent-validity evidence. AI acceptance was positively associated with intercultural teaching self-efficacy and autonomous teaching motivation. Self-efficacy was associated with both occupational wellbeing and lower emotional exhaustion, whereas autonomous motivation was associated primarily with occupational wellbeing. Direct associations between AI acceptance and the occupational-health outcomes were not significant. Conclusion The present findings provide preliminary evidence for using the overall AITA score as a context-specific measure of AI acceptance in intercultural teaching. The results also show cross-sectional associations between AI acceptance and teacher-resource variables, but they do not establish causal effects, mediation, or intervention effectiveness. The four TAM-based dimensions should be treated as provisional content facets requiring further validation.