Self-Regulated Learning with AI: A Comparative Analysis of General-Purpose and Task-Specific Platforms

Artificial intelligence (AI) offers opportunities for enhancing student self-regulated learning (SRL). This study investigates how two types of AI platforms—general-purpose AI (e.g., ChatGPT) and task-specific AI (e.g., EduGPT)—support SRL and satisfy students’ psychological needs. Grounded in Zimmerman’s SRL model and Self-Determination Theory (SDT), we examine the cognitive and motivational affordances provided by each AI type across multiple SRL phases. An experimental design involving 258 undergraduate students was implemented over an eight-week period. Participants were divided into three groups: general-purpose AI, task-specific AI, and control group. MANOVA results revealed that general-purpose AI tools primarily supported higher levels of autonomy and encouraged SRL skills such as goal setting, metacognitive reflection, and independent problem-solving. In contrast, task-specific AI tools were more effective in fostering competence and relatedness by providing structured guidance, timely feedback, and opportunities for social interaction, thereby enhancing effort regulation and social support. Thematic analysis further demonstrated distinct patterns in SRL strategies, with general-purpose AI promoting flexible self-directed learning, while task-specific AI provided scaffolding that encouraged incremental skill-building and collaboration. These findings underscore the complementary roles of the two AI tools in educational contexts, suggesting that a hybrid approach may optimize SRL.