Exploring Non-English Majors’ Perceptions Towards ChatGPT in Enhancing Speaking Accuracy: A Mixed-methods Study in HUIT

Abstract

This study investigated non-English major sophomores’ perceptions of using ChatGPT to enhance speaking accuracy (SA) at Ho Chi Minh City University of Industry and Trade (HUIT). Conducted over eight weeks with 137 participants, and based on the Technology Acceptance Model (TAM), the research used mixed-methods quasi-experimental design to have a deeper insight about how students perceived ChatGPT’s ease of use and usefulness in developing SA in terms of pronunciation, grammar and vocabulary. The findings revealed a slight enhancement in post-test compared to pre-test, particularly in vocabulary and grammar. Most students perceived ChatGPT is user-friendly and beneficial for language learning; however, it was suggested to be a supporting tool rather than a substitution for human teachers. The findings highlight promises of AI tools to assist speaking development in EFL contexts, offering practical implications for educators, universities and EdTech developers.