Exploration of University Students' Use of Artificial Intelligence for Independent Practice of English Speaking

This study investigated the extent of student engagement with an AI-powered speaking app in a self-directed English-speaking practice setting, focusing on the relationship between usage time and second language Willingness to Communicate (L2 WTC) in and outside classrooms. Conducted at a key Chinese university, the research utilized a mixed-methods approach with surveys, usage data, and study logs from 48 sophomore students, who had approximately 10 years of English learning experience. The AI tool, TalkEzly (powered by ChatGPT 3.5), was introduced with a self-regulation framework to guide students’ independent practice. Findings revealed that only half of the students engaged actively with the AI tool, with moderate users showing a significant increase in L2 WTC. Key perceived benefits included improved speaking skills, enhanced confidence, and vocabulary expansion, though a few students reported no notable improvements. This suggests that the app's effectiveness is linked to individual engagement. While the study underscores the AI app's potential to complement classroom instruction and foster independent language learning, it also highlights the need for sustained support to promote active usage. Future research should explore strategies to support student usage, verify the relation between the increase in L2 WTC and moderate usage, and examine the long-term impacts of AI tools on L2 development. As AI technology continues to evolve, educators must adapt to optimize its integration in language learning.