This review synthesizes theoretical frameworks, empirical findings, and methodological advancements in the domain of AI-driven self-regulated learning (SRL) within university-level English education. Grounded in social cognitive theory and extended through innovative models such as the "compound brain" paradigm, the integration of artificial intelligence (AI) technologies fosters dynamic human-machine collaboration to support multifaceted SRL processes, including goal-setting, strategy use, time management, self-evaluation, environment construction, and help-seeking. Empirical studies highlight robust measurement and classification techniques for SRL abilities, incorporating psychometric scales, latent profile analyses, and coding schemes based on learning analytics data. Generative AI tools like ChatGPT show promise in scaffolding autonomous English learning, particularly via writing assistance and adaptive feedback, though their efficacy is moderated by learner cognitive and meta-cognitive readiness. The interplay of motivation, positive psychology constructs, and meta-cognitive strategies emerges as a critical influence on SRL engagement and outcomes. Furthermore, the review delineates the process characteristics of SRL within hybrid and AI-supported online learning contexts, illuminating how peer collaboration and AI mediation interactively shape learning regulation. Despite notable progress, challenges persist in terms of conceptual clarity, theoretical diversification, methodological rigor, learner variability, and ethical considerations, notably regarding data privacy and educational equity. Future research trajectories advocate for integrative theoretical frameworks, expanded empirical samples, longitudinal designs, enriched meta-cognitive interventions, and systematic ethical analyses to deepen understanding and optimize AI-supported SRL in EFL contexts.
Chen Maochun (Wed,) studied this question.