Developing and Examining the Responsible AI Use Knowledge Scale (RAUKS): Initial Evidence of a Three-Factor Structure among College Students in Teacher Education

As generative artificial intelligence (GenAI) becomes increasingly embedded in online and blended higher education, pressing assumptions about authorship, academic integrity, and transparency are being re-examined, raising urgent questions for the transformation of the teaching and learning environment. While existing studies document students’ attitudes toward AI and general AI literacy, fewer instruments assess students’ self-reported understanding of responsible academic AI use. This study reports the development and preliminary validation of the Responsible AI Use Knowledge Scale (RAUKS) for college students in teacher education. Guided conceptually by the 6Cs approach as an item-development heuristic, an initial 46-item pool was administered to 410 undergraduate students. Exploratory structural equation modeling (ESEM) supported a refined 24-item instrument with a three-factor structure: Ethical Normative Knowledge (ENK), Critical Corrective Knowledge (CCK), and Procedural Disclosure Knowledge (PDK). The three-factor solution explained 61.67% of the total variance and demonstrated excellent model fit (χ² (230) = 364, p < .001, RMSEA = .038, SRMR = .033, CFI = .999, TLI = .999). Internal consistency reliability was strong across factors (ω = .97 for ENK, .84 for CCK, and .81 for PDK). The results provide initial evidence that responsible AI use perceived knowledge is multidimensional, comprising integrity-related expectations, critical evaluation of AI outputs, and disclosure procedures. RAUKS contributes a knowledge-focused self-report assessment tool offering practical utility for diagnosing student preparedness, informing instructional design, and supporting integrity-oriented AI policies. Further cross-validation across institutions, programs, and modalities is recommended.