<![CDATA[The growing global demand for English proficiency has coincided with the increasing importance of technological literacy, accelerating the integration of artificial intelligence (AI) in education. This study investigates the effectiveness of ELSA Speak, an AI-powered language learning application, in enhancing students’ speaking proficiency. A quasi-experimental mixed-method approach was employed, involving both an experimental and a control group. The research was conducted at Universitas Bina Sarana Informatika, where only two classes were available, prompting the use of convenience sampling to select participants based on availability and willingness. The sample included 35 students from class 33.1A.07 (experimental group) and 36 students from class 33.1B.07 (control group). Quantitative data were collected through oral pre-tests and post-tests and analyzed using t-tests in SPSS version 25. Qualitative data from in-depth interviews provided insights into how features such as structured levels and personalized feedback influenced students’ speaking development. Results showed that the control group’s scores increased from 58.78 to 65.81, while the experimental group’s scores rose from 61.43 to 75.03. Statistical analysis revealed a significance level (2-tailed) of <0.05, leading to the rejection of the null hypothesis and acceptance of the alternative. The findings indicate that ELSA Speak significantly enhanced students’ speaking proficiency in grammar, vocabulary, pronunciation, fluency, and content. The study concludes that integrating AI-based tools like ELSA Speak into university curricula offers a practical solution for improving language skills, especially in institutions with limited resources and instructional time]]>
