Reconsidering the role of L2 self-motivational and affective factors in AI-mediated informal digital learning of English: a mixed-methods study

Introduction: Growing scholarly interest has emphasized the significance of non-cognitive variables in influencing learners' informal English acquisition through artificial intelligence (AI)-enhanced digital platforms. Despite this, research remains sparse on how motivational constructs, emotional responses, particularly enjoyment, and anxiety resulting from AI-assisted learning (AAL), and participation in AI-mediated informal digital English learning (AI-IDLE) collectively contribute to learners' self-efficacy beliefs in speaking English as a foreign language (EFL). Methods: Employing an explanatory sequential mixed-method design, this study surveyed 308 EFL secondary students in China, followed by in-depth interviews with eight voluntary respondents. Quantitative data were analyzed using structural equation modeling, while qualitative data underwent thematic analysis involving both open and axial coding procedures. Results: Path results demonstrated that AAL negatively predicted EFL learners' AI-IDLE engagement. However, AI-IDLE exerted no statistically significant influence on students' self-reported self-efficacy in spoken English. Notably, enjoyment emerged as a mediator in the paths from the ideal second language (L2) self to both AI-IDLE and self-efficacy beliefs in speaking capacity. Qualitative data based on narrative descriptions evinced the heterogeneity of motivational and emotional factors of students operating in AI-supported informal learning situations. Discussion: Through a combination of quantitative results with interpretative reflections, this study presents some useful guidelines for maximizing student participation with respect to AI-IDLE, ensuring AI literacy, and protecting their emotional wellbeing through the rapidly developing environment of generative AI.