<![CDATA[Artificial Intelligence (AI) is now beginning to be integrated into language learning, particularly in pronunciation practice. However, there is a lack of research exploring students' perceptions and challenges in using AI-based applications for pronunciation practice, especially in the context of English as a foreign language (EFL) learning in Indonesia. This study aims to investigate students' perceptions and challenges toward the use of AI-based applications for pronunciation practice by applying the Technology Acceptance Model (TAM) by Davis (1989) and Challenges with the Use of AI in ELT/L framework by Crompton et al. (2024). The research employed a qualitative study with data collection through questionnaires and semi-structured interviews involving 34 English Language Education students. The study found that students view AI-based applications, specifically Google Translate, Elsa Speak and Duolingo, as effective and user-friendly tools that provide personalized feedback, repetitive practice, and flexible learning plans. Apart from that, students also feel that the apps help them to gain a better understanding about important features like consonants, intonation and stress. However, several challenges were identified, including technical problems, recognition errors, privacy concerns, and doubts about the accuracy of feedback. Despite these challenges, AI-based applications show promising potential to support pronunciation learning if there are further improvement to address existing limitations.]]>
