AI Chatbots in ELT: A systematic review of pedagogical use and challenges in higher education

The advancement of artificial intelligence (AI) technology has significantly impacted the field of education, particularly in English Language Teaching (ELT) at the higher education level. One prominent innovation is the use of AI-based chatbots as interactive learning partners. This study aims to systematically examine the pedagogical roles, potentials, and implementation challenges of AI chatbots in the context of ELT in Indonesian higher education institutions. Employing a Systematic Literature Review (SLR) approach, this study analyzes 56 articles published in nationally indexed journals (SINTA 1 to 6) between 2020 and 2025. The analysis is grounded in the Technological Pedagogical Content Knowledge (TPACK) and Second Language Acquisition (SLA) frameworks, focusing on three main domains of challenges: technical, pedagogical, and affective. Findings indicate that AI chatbots such as ChatGPT, ELSA Speak, and Gemini play pivotal roles as writing assistants, intelligent tutors, and automated evaluation tools in English language instruction. The use of chatbots has been shown to improve students' writing and speaking skills, support autonomous learning, and enhance motivation and engagement. However, implementation still faces significant challenges, including infrastructure limitations, low digital literacy among instructors, risks of plagiarism, and cultural resistance to technology. The study recommends continuous technical and pedagogical training, strategic integration of AI into the curriculum, and the development of ethical and regulatory policies to ensure the effective and responsible use of AI in higher education.