How AI literacy predicts L2 Chinese writing engagement: the mediating roles of anxiety and enjoyment.

Introduction: In the era of Generative AI, the focus of second language (L2) writing research has shifted toward how learners' AI literacy shapes their affective experiences and engagement. Drawing upon the Control-Value Theory (CVT), this study investigated the mechanisms underlying the relationship between L2 Chinese Writing AI Literacy and L2 Engagement, with L2 Writing Anxiety and L2 Writing Enjoyment as mediating variables. Methods: The study employed an explanatory sequential mixed-methods design. In the quantitative phase, structural equation modeling (SEM) was conducted with 408 L2 Chinese learners. In the subsequent qualitative phase, Stimulated Recall Diaries (SRD) and Stimulated Recall Interviews (SRI) were conducted with eight participants to gain an in-depth understanding of their emotional trajectories. Results: Quantitative results confirmed that L2 Chinese Writing AI Literacy significantly predicted L2 Engagement both directly and indirectly. Specifically, AI literacy served as a powerful affective regulator, significantly reducing L2 anxiety by enhancing learners' perceived control and thereby boosting writing engagement. Qualitative findings revealed that AI literacy empowered learners to transition from anxious novices to confident editors, where the exercise of selective agency became a primary source of control. Discussion: Overall, the study highlights the role of L2 Chinese Writing AI Literacy as an emotional buffer for creative agency in technologically enhanced writing environments. The findings offer critical pedagogical implications for fostering AI literacy and developing differentiated scaffolding strategies to optimize L2 Chinese writing development.