The Role of Adaptive Artificial Intelligence Education on Students' Deep Learning with the Mediation of Academic Self-Efficacy

Abstract

The purpose of the present study was to investigate the mediating role of academic self-efficacy in the relationship between the perception of adaptive artificial intelligence-based education and students' deep learning. The research method was applied in terms of purpose and descriptive-correlational in terms of nature. The statistical population included all first-grade secondary school students in the Serishabad region in the 2024-2025 academic year. Using Cochran's formula, 168 individuals were selected via simple random sampling. To collect data, three questionnaires were used: a researcher-made questionnaire on the perception of adaptive AI education, the Jinks and Morgan (1999) academic self-efficacy questionnaire, and the Biggs et al. (2001) learning approaches questionnaire. The face and content validity of the instruments were confirmed by experts, and their reliability was reported as satisfactory using Cronbach's alpha coefficient. Data were analyzed using SPSS software at both descriptive and inferential levels (Pearson correlation and multiple regression). The findings showed that there is a positive and significant relationship between adaptive AI education and academic self-efficacy as well as deep learning. Furthermore, the regression analysis results indicated that academic self-efficacy plays a partial but significant mediating role in this relationship; meaning that adaptive AI education leads to the enhancement of students' deep learning both directly and by strengthening self-efficacy. This study emphasizes that utilizing intelligent technologies, with an emphasis on psychological dimensions, can provide a suitable foundation for realizing deep learning.