Moderating the Effects of Artificial Intelligence on Language Learning: Prior Knowledge and Psychological Needs Satisfaction in Self‐Regulated Learning

ABSTRACT This study examines how artificial intelligence (AI) tools support self‐regulated learning (SRL) and psychological need satisfaction in a Chinese university English‐as‐a‐foreign‐language (EFL) context, focusing on the moderating role of prior knowledge, operationalized in this study as baseline English proficiency. Using a quasi‐experimental design, 190 undergraduates enrolled in a mandatory College English course engaged with either general‐purpose AI tools offering flexible, open‐ended affordances or specialized AI tools providing structured, task‐specific support. A baseline proficiency pretest and post‐intervention self‐report measures were used to assess prior knowledge, SRL skills, and satisfaction of autonomy, competence, and relatedness. Moderated multiple regression analyses revealed significant interactions between baseline English proficiency (prior knowledge) and AI affordance type. For learners with higher baseline proficiency, general‐purpose AI tools were more effective, strengthening SRL and psychological need satisfaction through exploratory and customizable support. For learners with lower baseline proficiency, specialized AI tools provided beneficial scaffolding but were less supportive for advanced learners by constraining autonomy and flexibility. These findings extend SRL and Self‐Determination Theory (SDT) by specifying when different AI affordances are likely to facilitate or hinder SRL processes and need satisfaction. The study highlights the importance of aligning AI affordances with learner characteristics and offers context‐sensitive design implications for adaptive AI and tool selection in L2 classrooms.