Speaker
Description
Speaking is often the most challenging skill for EFL learners in rural Vietnam due to limited access to authentic English input, teacher support, and learner confidence. This action research explores the use of an AI-powered speaking platform developed by IMME Australia, modeled after the Pearson Test of English (PTE), to improve students’ oral performance. The study involved 40 lower secondary students (aged 13–15) in a rural Vietnamese school over eight weeks. A mixed-methods approach was adopted, including pre- and post-assessments, classroom observations, and student surveys and interviews. Findings revealed significant improvements in pronunciation, fluency, and speaking confidence. Students appreciated practicing with the AI tool, noting reduced anxiety and increased comfort speaking without peer judgment. Key features such as accent tolerance, instant feedback, and repeatable practice were especially helpful for shy or low-performing students. The AI platform served not as a replacement for human interaction, but as a valuable supplement that created a safer, more equitable speaking environment. This presentation shares implementation strategies, student reflections, and practical recommendations for integrating AI tools in low-resource EFL contexts. It also discusses implications for future speaking instruction and assessment in the age of AI, particularly in underserved regions like rural Vietnam.
Keywords: AI in language learning, pronunciation, speaking confidence, rural Vietnam
