Presentation The Feedback Loop: How ESL Learners Navigate Autonomy and Trust in ChatGPT Interactions more

Generative artificial intelligence (GenAI) is rapidly transforming English language learning by offering personalized, interactive support for vocabulary development (Moorhouse et al., 2025). Yet, while linguistic outcomes are increasingly studied, less is known about how learners experience these tools (Zhang, 2025). This study investigates the ways university‑level ESL students interact with generative AI and how this interaction influences their motivation, confidence, and sense of autonomy. 87 undergraduate ESL learners, mainly L1 Japanese speakers with some Chinese and Korean participants, used ChatGPT-4o in a medical vocabulary course. Two groups completed identical tasks: one received English-only feedback, while the other utilized their first language (L1) for feedback and prompt adaptation (Moorhouse et al., 2025; Yang & Lin, 2025). Data from surveys and interviews captured perceptions, usability challenges, and learning outcomes. Findings indicate that GenAI boosts motivation through immediate, contextualized vocabulary explanations (Xiao & Zhi, 2023). However, learners faced limitations like repetitive feedback (Yang et al., 2025) and unreliable voice recognition, which affected trust (Mompean, 2024). The study highlights integrating GenAI within pedagogically scaffolded instruction (Park & Kim, 2025). When guided effectively, AI tools promote reflective learning, critical engagement, and greater learner autonomy (Park & Kim, 2025). These insights help TESOL practitioners understand how GenAI can be meaningfully incorporated into communicative vocabulary teaching.