Rahma, Riza Aliyata. (2024). The Effectiveness of Using ChatGPT on Students’ Reading Comprhension Skills. Thesis. Department of English Language Education, Faculty of Tarbiyah, State Islamic Institute (IAIN) of Kediri. Advisors: (I) Bahruddin, S.S., M.Pd. (II) Dr. Fathor Rasyid, M.Pd. Keywords: Artificial Intelligence, ChatGPT, Students’ Achievement. Language is a tool for communication, and English as a global language is widely taught from kindergarten to higher education, depending on each country’s educational system. The implementation of English language learning in has evolved, with modern educational practices and multimedia integration. The integration of digital tools and AI systems, such as ChatGPT, has made learning more engaging and effective. However, there are challenges in adapting these technologies to traditional classroom, as the researcher observed at MTs Miftahul Huda Madiun, where the teachers still use textbook-based learning. Therefore, this study is conducted in order to explore the effectiveness of ChatGPT to improve students’ English achievements, specifically on reading comprehension skills. This study used quantitative research methodology with quasi-experimental design to answer the purpose of the research. The subjects of this research were 8th-grade students of MTs Miftahul Huda Madiun. The research instrument used to collect the data is test, given by the researcher before and after the treatment. To analyze the data, the researcher used statistical analysis of ANCOVA using SPSS 25. The studys results were analyzed using ANCOVA with SPSS 25. The Shapiro-Wilk test indicated normality, with p-values of 0.357 (pre-test) and 0.294 (post-test) for the experimental group, and 0.068 (pre-test) and 0.529 (post-test) for the control group, confirming normal distribution since all were above 0.05. Levene’s test showed a significance of 0.468, indicating homogeneous variances. The interaction between group and pretest was 0.290, meaning no interaction and fulfilling the homogeneity of regression assumption. The covariates significance was 0.004, indicating a linear relationship with the dependent variable. Finally, ANCOVA showed a significance value of 0.001 for the groups, leading to the rejection of H₀ and acceptance of H₁, demonstrating a significant difference in reading comprehension improvement between students using ChatGPT and those using traditional textbooks.

RAHMA , RIZA ‘ALIYATA .   (2025). THE EFFECTIVENESS OF USING CHATGPT TO IMPROVE STUDENTS’ READING COMPREHENSION SKILLS. IAIN Kediri