Artificial Intelligence–Based English Instruction: Implementation, Challenges, and Learning Outcomes in Indonesian High Schools

Abstract

This study aims to examine the implementation, challenges, and learning outcomes of Artificial Intelligence (AI)-based English instruction in three Indonesian Islamic senior high schools. Employing a descriptive qualitative research design, the study explores students’ perceptions, learning behaviors, and interactions with AI tools such as ChatGPT, Grammarly, Google Translate, and AI-driven pronunciation systems. Data were collected through open-ended questionnaires and analyzed using thematic procedures. Findings indicate that AI tools contribute positively to students’ vocabulary development, grammatical awareness, writing fluency, and speaking confidence by offering immediate feedback, adaptive support, and opportunities for autonomous learning. Nonetheless, students also experienced key challenges, including misinformation, cognitive overload from complex explanations, overdependence on AI-generated responses, and unequal access to technological resources. These constraints influenced the extent to which AI could be integrated effectively into English instruction. The discussion highlights that although AI serves as a valuable supplementary learning resource, it cannot replace the pedagogical, emotional, and ethical roles fulfilled by teachers. Students emphasized the need for human guidance to interpret AI-generated information, ensure accuracy, and maintain academic integrity. The study concludes that AI-enhanced English learning holds substantial potential when implemented responsibly. It underscores the importance of developing AI literacy, strengthening digital infrastructure, and establishing ethical guidelines to support sustainable integration of AI in Indonesian secondary education.