This study examined the role of learning engagement and academic self-efficacy in the feelings of burnout among English as a foreign language (EFL) students from China (N = 197). Three closed-ended scales were employed in the quantitative phase of this study to explore the relationship between the latent variables. Additionally, a hypothetical model was created and verified using Pearson's correlation analysis, structural equation modeling (SEM), path analysis, and mediating effect test. The investigation results demonstrated that while learning engagement was positively correlated with medium levels of academic self-efficacy, student burnout was inversely correlated with both variables. Among Chinese EFL students, academic self-efficacy and learning engagement are predictive of burnout, and learning engagement can act as a mediator to reduce the impact of academic self-efficacy on burnout. The implications for EFL teaching and learning are discussed in light of those findings, including enhancing EFL students' self-efficacy to reduce pessimism and increase academic achievement.

The rapid integration of artificial intelligence (AI) into language education has transformed how learners manage and direct their learning processes. Despite the growing adoption of AI-assisted tools, empirical understanding of their psychological and behavioral impacts remains incomplete. This study investigated how engagement in AI-assisted language learning environments shapes learners’ self-regulation, autonomy, and self-directed learning behaviors. Drawing on self-determination theory and Zimmerman’s model of self-regulated learning, the research employed a quantitative design with data collected from 736 Chinese university students using validated questionnaires measuring AI engagement, self-regulation, self-directed learning, and learner autonomy. Structural equation modeling (SEM) and correlational analyses were conducted using SPSS (v27) and AMOS (v24). Results indicated strong positive correlations between AI engagement and self-regulation, self-directed learning, and autonomy. Moreover, AI engagement significantly predicted learners’ self-regulation and autonomy, whereas self-regulation partially mediated the relationship between AI engagement and self-directed learning. These findings suggest that AI technologies, when employed as autonomy-supportive tools, can strengthen learners’ metacognitive awareness, intrinsic motivation, and independence in language learning. The study offers theoretical insights into digital self-regulated learning models and provides practical implications for educators seeking to integrate AI systems in ways that foster sustainable learner autonomy.