Speaker
Description
ChatGPT has been a commonly used tool for facilitating speaking practice and feedback, its contribution to improving vocabulary complexity has yet to be adequately examined and evaluated in an educational context.
This quasi-experimental study was carried out at OVA English Center in Vietnam, where 30 IELTS learners, whose initial speaking level ranged from Band 5.0 to 5.5, were randomly assigned to either the control or experimental groups. Before the study, an IELTS speaking pre-test was administered to all participants. For eight weeks, both groups were given the same IELTS speaking topics. The control group was given traditional peer-speaking practice and peer-evaluation tasks, while the experimental group used their mobile devices to access ChatGPT and were provided with immediate lexical feedback on their speaking performance. At the end of the study, an IELTS speaking post-test was administered to both groups. The speaking performances were recorded, transcribed, and graded in line with the official IELTS Lexical Resource criteria. Vocabulary associated with specific topics is categorized and analyzed statistically by applying paired t-tests and independent samples t-tests to compare and identify any differences between and within both groups.
In examining Lexical Resource as a quantified variable, this study aims to highlight AI’s role in education and offers experimental results from actual educational contexts regarding the potential for teaching with ChatGPT. The findings aim to guide new and creative approaches to IELTS instruction and AI integration with Vietnam’s emerging ESL environment.
Keywords: ChatGPT, Lexical Resource, Topic-specific vocabulary, IELTS Speaking, AI-assisted learning, ESL classroom innovation
