<p>This study was an endeavor to understand the factors driving first-year university students’ engagement in AI-mediated informal English learning (AI-IDLE). To do so, we formulated a conceptual model by integrating the capital theory and the task-technology fit theory to explore whether and how cultural capital (i.e., AI literacy, techno-growth mindset, and technostress), social capital (i.e., facilitating conditions and social influence), intrinsic motivation (i.e., perceived enjoyment), and task-technology fit influence students’ learning engagement in AI-IDLE. A sample of 686 first-year university students was collected from four public universities in South China using a self-reported survey. The results of structural equation modeling show that the positive effect of social influence on AI-IDLE engagement is the greatest, followed by AI literacy, perceived enjoyment, and techno-growth mindset. However, facilitating conditions do not impact AI-IDLE engagement significantly. Additionally, technostress decreases the positive effect of social influence on learning engagement, while task-technology fit increases the positive effect of perceived enjoyment on learning engagement. This study expanded the literature on AI embracement by exploring the factors influencing newly enrolled undergraduates’ self-directed English learning. The findings suggested that universities should take effective measures to build a positive AI-mediated English learning atmosphere, cultivate first-year students’ AI literacy, and techno-growth mindset.</p>