Lai, Zola

Center for English Language Teaching, Wenzao Ursuline University of Languages

Abstract: L2 listening requires learners to manage attention, monitor understanding, and recover when comprehension breaks down. As AI tools enter listening classrooms, a key question is not simply whether AI helps, but how different forms of support relate to learners’ reported metacognitive awareness. This study compared AI guided support, AI flex use, and no AI instruction in a textbook-based listening course with 109 Taiwanese university EFL learners.
The overall MALQ results showed that the AI guided group had the highest adjusted mean, followed by the no AI group and then the AI flex group, although only the AI guided versus AI flex contrast was statistically significant. The subscale results told a more specific story. AI guided learners reported stronger awareness of planning and evaluation and directed attention, and showed the lowest self-reported reliance on mental translation. No AI learners reported stronger problem-solving awareness, suggesting that peer discussion and teacher-mediated support may still matter when learners work through listening difficulties. Learner reflections suggested that structured AI support provided students with clearer guidance for approaching listening, whereas flexible AI use often left them uncertain about when to seek help.

Key words: AI-assisted listening, AI support configurations , L2 listening , metacognitive awareness

Note: This is a pre-proof version and is subject to change during the editing process.