Development and validation of a multidimensional scale for AI tutoring acceptance in higher education

Artificial intelligence (AI) is being increasingly applied in higher education. This has given rise to the issue that, within the published literature, there is still a lack of widely established and psychometrically validated tools for assessing the actual perceived acceptance quality of AI tutoring systems in a comprehensive manner. Although previous studies have examined cognitive patterns in educational technology, few have adequately captured the multidimensional nature of learners' experiences with AI tutoring systems. Existing evaluation models mainly emphasize cognitive outcomes, with less attention given to motivation, relational experience, and trust in AI-assisted teaching contexts. This study aimed to develop and validate a learner-centered scale of AI tutoring acceptance quality. The scale was constructed through literature review, expert consultation, and pre-testing, and comprised four theoretically grounded dimensions: Feedback Quality, Instructional Effectiveness, Interaction Experience, and User Trust. Data were collected from undergraduate students at five universities in Shanxi Province, China. ESEM, bifactor modeling, and reliability analysis indicated a robust bifactor structure, good internal consistency, and acceptable construct validity. The validated scale may serve as a practical tool for AI-supported educational design and future intervention research.