This study examines how students' English learning motivation changes within an AI-based personalized learning environment and identifies the differentiated characteristics that emerge in this process. Four low-achieving students from a specialized vo...
This study examines how students' English learning motivation changes within an AI-based personalized learning environment and identifies the differentiated characteristics that emerge in this process. Four low-achieving students from a specialized vocational high school participated in a 30-day English learning program using READING &, which integrates three AI learning support features: personalized learning system, real-time feedback, and gamification. Data were collected through semi-structured interviews and learning logs. Findings show that students displayed distinct appropriation patterns depending on their initial motivational states, despite engaging with the same AI features. These patterns included trusting and accepting system guidance, adjusting features to align with personal goals, critically evaluating AI feedback, or reorganizing features for alternative uses. These appropriation processes shaped individualized patterns of motivational improvement and revealed a cyclical interaction between technological structures and students' appropriation behaviors. Particularly, relatedness needs were experienced in expanded digital forms through repeated interactions with AI, manifested as psychological safety, a sense of personalized learning identity, and participation in a virtual learning community, with AI functioning as an alternative or complementary support network. Moreover, the same technological structure functioned as ‘structural guide’, ‘efficiency optimizer’, ‘exploratory learning space’, or ‘relational learning hub’ depending on students' initial motivational states, creating differentiated pathways for motivational improvement. The study integrates self-determination theory (SDT) and adaptive structuration theory (AST) to explain these mechanisms and systematically confirms a cyclical structure where initial motivation influences appropriation patterns, which in turn reshape motivation. Based on these findings, the study underscores the importance of adaptive design based on students' prior motivational profiles and synergistic integration of AI learning support features, suggesting that AI-based learning motivation research should be expanded toward an integrative theoretical framework that captures the dynamic interaction between technology and learners.