Development and Validation of a Multidimensional Gen-AI Competence Scale for Pre-Service Science Teachers

The rapid integration of generative artificial intelligence (Gen-AI) into education is transforming the competencies required of future teachers and creating a need for a robust multidimensional measure of Gen-AI competency. This study developed and validated a six-dimensional framework comprising output evaluation, use appropriateness, prompting & operation, technical integration, model literacy, and responsible practice among pre-service science teachers. Data were obtained from 426 biology, chemistry, and physics education students at Universitas Jambi, Indonesia. The responses were randomly split into EFA (n = 256; 60%) and CFA (n = 170; 40%) subsamples. An initial 84-item pool was evaluated through ten experts’ judgment using CVR and CVI, resulting in 63 content-valid items. The data were suitable for factor analysis (KMO = 0.972; Bartlett’s χ² = 20430.533; df = 1,953; p < 0.001), and EFA retained 33 items across six dimensions, with factor loadings of 0.502–0.811 and Cronbach’s alpha of 0.765–0.909. CFA supported the six-dimensional structure with acceptable fit (RMSEA = 0.063, CFI = 0.913, TLI = 0.904, SRMR = 0.052). The dimensions demonstrated satisfactory reliability and convergent validity (α = 0.754–0.909; CR = 0.752–0.910; AVE = 0.505–0.611), while HTMT values of 0.586–0.874 supported discriminant validity. The resulting measure provides an empirically supported basis for profiling Gen-AI competency and strengthening teachers’ preparation for responsible Gen-AI use.