Saudi EFL Learners’ Perceptions of AI Chatbots for Speaking Development: A Qualitative Study of ELSA Speak

Abstract

Aims: This study investigates Saudi EFL learners’ perceptions of the AI-powered chatbot ELSA Speak in developing speaking proficiency, specifically fluency and pronunciation accuracy, and the challenges encountered during chatbot use. Study Design: A qualitative research design grounded in an interpretivist paradigm was employed. Place and Duration of Study: The study was conducted at King Abdulaziz University, Jeddah, Saudi Arabia, during the first semester of the 2024–2025 academic year, over a three-week chatbot trial period. Methodology: Data were gathered from eight female undergraduate preparatory-year students through semi-structured interviews conducted after the trial period and analysed using Braun and Clarke’s (2006) six-phase thematic analysis. Ethical approval was obtained, and written informed consent was secured from all participants prior to data collection. Results: Four themes emerged: fluency enhancement through a permanently available speaking partner and uninterrupted practice; improved pronunciation accuracy through immediate corrective feedback and clear modelling; curriculum disconnect due to misalignment between chatbot content and academic objectives; and feedback challenges arising from language complexity beyond learners’ proficiency level. Conclusion: AI chatbots hold considerable potential for developing EFL learners’ speaking skills, particularly where opportunities for oral practice are limited. However, effective integration requires alignment with institutional curricula and calibration to learners’ proficiency levels, suggesting the need for institution-specific AI tools designed to meet curricular objectives and learner needs.