Introduction The proliferation of generative artificial intelligence (AI) has created unprecedented opportunities for informal second language (L2) learning, yet the psychological mechanisms through which AI-mediated learning translates into communicative engagement remain poorly understood. This study examines how AI-mediated informal digital learning of English (AI-IDLE) shapes learners' willingness to communicate (WTC) in an L2 through the dimensional specificity of foreign language enjoyment (FLE). Methods A large-scale investigation was conducted with 1,362 non-English-major university students in China. Using structural equation modeling with confirmatory factor analysis, parallel mediation models were estimated in which receptive and productive AI-IDLE predicted in-class and out-of-class L2 WTC through three FLE dimensions simultaneously: teacher appreciation, personal enjoyment, and social enjoyment. Results FLE partially mediated the AI-IDLE-WTC link, but the three FLE dimensions operated through context-dependent pathways. Personal enjoyment emerged as a robust cross-contextual mediator in both classroom and extracurricular settings, whereas teacher appreciation mediated the relationship exclusively in structured classroom environments and showed no significant indirect effect in out-of-class contexts. Social enjoyment contributed to mediation across both settings, but with markedly attenuated strength outside the classroom. Productive AI-IDLE exhibited stronger associations with all FLE dimensions than receptive AI-IDLE. Discussion These findings advance a dimensional specificity framework for understanding how positive emotions differentially channel technology-enhanced learning into communicative behavior. The results carry implications for the design of AI learning tools that deliberately target distinct affective pathways and underscore the importance of context-sensitive approaches to fostering L2 willingness to communicate.
Guo et al. (Mon,) studied this question.