THE IMPACT OF USING THE VOICE CONVERSATION FEATURE IN CHATGPT ON SPEAKING SKILLS AND SPEAKING ANXIETY AMONG IRAQI UNIVERSITY STUDENTS

Objective: This study investigated the impact of using the voice conversation feature in ChatGPT on speaking skills and speaking anxiety among Iraqi university students. Method: A quasi-experimental design was employed with 70 second-stage students from the Department of English, Faculty of Education, University of Kufa. Participants were assigned into experimental (n=35) and control (n=35) groups. The experimental group practiced speaking with ChatGPT's voice feature for six weeks (15 minutes in class and 15 minutes daily at home), while the control group received no treatment. Data were collected using a speaking skills test and a speaking anxiety scale adapted from Horwitz et al. (1986). Results: Results of independent samples t-tests revealed that the experimental group significantly outperformed the control group in speaking skills (t=10.847, p<0.001, d=2.59) and showed significantly lower speaking anxiety (t=-11.034, p<0.001, d=-2.64). Paired samples t-tests showed significant pre-post improvements in the experimental group for speaking skills (mean difference=+20.5, t=15.342, p<0.001) and anxiety reduction (mean difference=-24.3, t=-14.876, p<0.001), while the control group showed no significant changes. Novelty: These findings demonstrate that ChatGPT's voice conversation feature is an effective tool for improving speaking skills and reducing speaking anxiety among Iraqi university students. The study provides practical implications for integrating AI tools into English language instruction in Iraqi higher education.