The rapid development of artificial intelligence has created new opportunities to improve English language instruction, especially in speaking abilities. Nevertheless, eleventh-grade learners at MAN 2 Bukittinggi continue to face challenges in English speaking, such as limited vocabulary knowledge, lack of self-assurance, and inadequate opportunities for speaking practice. Moreover, the teaching of English speaking at MAN 2 Bukittinggi predominantly continues to rely on traditional methods, which may constrain students' chances to participate in meaningful and interactive speaking sessions. Consequently, The purpose of this study was to investigate how using ChatGPT Audio Chat affected the speaking abilities of eleventh-grade students at MAN 2 Bukittinggi. A quasi-experimental two-group pre-test and post-test design was used in a quantitative manner. For the 2025–2026 school year, the sample was taken from all of the eleventh-grade students at MAN 2 Bukittinggi. Purposive sampling was used to choose the 32 participants, of which 16 students from class XI D1 were designated as the experimental group and 16 students from class XI D4 as the control group. The experimental group was instructed using ChatGPT Audio Chat, while conventional teaching methods were used for the control group. A speaking assessment was utilized, evaluated with an analytical rubric that focused on grammar, vocabulary, pronunciation, fluency, and content. Descriptive statistics, evaluations of homogeneity and normality, paired-samples t-tests, and independent-samples t-tests were all used in the data analysis, which was carried out using SPSS version 20. The experimental group's mean score increased from 45.63 in the pre-test to 58.38 in the post-test, according to the data, whereas the control group's scores decreased from 44.63 to 52.25. A significance level of 0.000 was found using the paired-samples t-test, indicating a significant improvement in the experimental group's students' verbal skills after the intervention. Furthermore, a significant level of 0.001 was shown by the independent-samples t-test, demonstrating a considerable disparity between the experimental and control groups. The experimental group also recorded a greater average score in the post-test compared to the control group. Thus, the implementation of ChatGPT Audio Chat positively influenced the speaking abilities of eleventh-grade students at MAN 2 Bukittinggi in a significant manner.
Copyrights © 2026