This study investigated the effects of AI conversational practice on university students’ speaking and listening skills, ChatGPT literacy, and perceptions of ChatGPT use across different English proficiency levels. Participants were divided into low- and high-proficiency groups. Both groups participated in oral practice with ChatGPT, targeting speaking and listening skills. Quantitative data were collected through pre- and post-listening and speaking tests and a 20-item ChatGPT literacy questionnaire that measured perceptions of usefulness, ease of use, affective responses, and strategic use of ChatGPT. Paired-samples t-tests, independent-samples t-tests, and ANCOVA were conducted to examine within-group gains and between-group differences. Qualitative data were obtained from openended survey responses and were analyzed using a thematic analysis. The results showed that both proficiency groups demonstrated statistically significant improvements in speaking and listening skills, although highproficiency learners exhibited larger gains in both areas. Both groups also showed statistically significant growth in ChatGPT literacy, and no statistically significant differences between proficiency levels were found. The findings highlight that learners at different proficiency levels derive distinct affective versus cognitive benefits from ChatGPT, underscoring the need for proficiency-sensitive instructional design and supporting its use as a supplementary tool to enhance oral proficiency and ChatGPT literacy.
Hea-Suk Kim (Sat,) studied this question.