AI competence domains and experience-based differences among science and technology teachers investigated using structural equation modeling

This study examined the structural relationships between domains of AI competence and age-adjusted differences based on computer science teaching experience among science and technology teachers. Grounded in Bandura’s Social Cognitive Theory, AI competence was conceptualized as a multidimensional configuration of interrelated domains relevant to instructional practice. A quantitative, nonexperimental, cross-sectional design was employed, with data collected from 1105 secondary school science and technology teachers in Thailand. Four domains of AI competence were examined: Human-Centered AI (HC), Ethics and Social Impact (ET), Knowledge and Applications (KN), and Classroom Implementation (CL). Structural equation modeling was used to examine structural associations among the AI competence domains, with model fit indices indicating acceptable to good fit (χ2/df = 46.50, CFI = 0.96, TLI = 0.95, RMSEA = 0.07, SRMR =0 .03). Repeated measures ANCOVA was conducted to examine age-adjusted differences in AI competence by computer science teaching experience. The results indicated that the four AI competence domains were structurally interconnected, forming an integrated configuration. The HC and ET domains were closely associated with conceptual understanding of AI, while Knowledge and Applications showed strong associations with CL. Teachers with greater computer science teaching experience reported higher levels of perceived AI competence across domains, whereas differences between adjacent experience groups were not statistically significant. These findings suggest that teachers’ AI competence is best understood as a structured configuration of interrelated domains, providing empirical evidence to inform future research and professional development in AI education.