Teaching Spoken English through Artificial Intelligence: A Systematic Review of Methods, Effectiveness, and Implementation Challenges (2015–2025)

The rapid advancement of artificial intelligence (AI) has transformed language education, particularly in the development of spoken English skills. This systematic literature review examines studies published between 2015 and 2025 on the use of AI in teaching spoken English, focusing on instructional methods, effectiveness, and implementation challenges. A comprehensive search using Google Scholar, followed by PRISMA-guided screening based on titles, abstracts, and keywords, resulted in 22 highly relevant studies from an initial pool of 671. The findings indicate that AI-driven technologies—especially conversational chatbots, automatic speech recognition (ASR) systems, generative AI models, and intelligent personal assistants—play a central role in facilitating interactive and learner-centered speaking practice. These approaches consistently improve pronunciation accuracy, oral fluency, and communicative competence. In addition, AI-mediated environments reduce speaking anxiety and develop learners’ willingness to communicate by providing low-pressure, non-judgmental interaction. However, several challenges persist, including ASR inaccuracies, infrastructural limitations, limited conversational depth, and insufficient pedagogical integration. The effectiveness of AI tools is also context-dependent, varying across learner proficiency and instructional design. Overall, AI functions most effectively as a complementary pedagogical partner, highlighting the need for longitudinal research and strategic integration in language education.