VietTESOL International Convention 2026

The growing availability of generative artificial intelligence (GenAI) tools has created new opportunities for extending language learning beyond the classroom. This action research study investigated the use of Microsoft Copilot as a speaking partner to improve students’ productive use and retention of newly learned vocabulary in a university level Listening and Speaking course. The seven-week study had students practice oral communication with Copilot outside of class using speaking prompts designed by the teacher. At the end of each lesson, students received a list of target vocabulary words and were asked to engage in spoken conversations with Copilot while blindfolded, promoting spontaneous language production while reducing reliance on written materials. Students then submitted transcripts of conversations and identified the target words that they had successfully used. Data were collected by means of transcript analysis and a post-intervention questionnaire. Findings revealed that students purposefully incorporated newly learned vocabulary into meaningful conversations and demonstrated improved retention through repeated use across different contexts. Questionnaire responses indicated positive attitudes toward the activity, with students reporting increased confidence, greater engagement, and enhanced awareness of vocabulary usage. The study suggests that structured AI-mediated speaking practice can effectively support vocabulary development while promoting meaningful and autonomous use of AI tools in language learning.