ENHANCING SPEAKING PROFICIENCY IN LANGUAGE LEARNING: HOW AI-ASSISTED GRAMMAR CORRECTION MEDIATES LEARNER CONFIDENCE AND FLUENCY

Abstract

The present study stems from an attempt to meet the pressing need within the second language acquisition (SLA) domain for innovative solutions to potentially improve speaking proficiency through an investigation into the intermediary nature of learners' confidence within the link between AI-facilitated grammar redress and speaking performance. The advancement of AI tools for language learning is evident, but it is not clear how this, in turn, affects the way learners perceive their confidence and fluency in spoken communication. Therefore, the purpose of this study is to provide a conceptual framework that illustrates how AI-assisted grammar correction builds student confidence and ultimately enhances speaking proficiency. Data were extracted from peer-reviewed articles available in the Scopus database utilizing a narrative review methodology, encompassing themes like personalized feedback, synchronous corrections, and sociocultural factors in AI-mediated language learning. The main findings are that AI tools reduce language anxiety, provide immediate and personalized feedback, and make learners more willing to engage in communication and lead to better speaking skills. The study highlights the need to consider contextual factors and potential overreliance by juxtaposing AI feedback with teacher input surrounding cultural relevance. Theoretically, this study presents Interaction Hypothesis and Self-Efficacy Theory to explain how AI tools use, confidence, and speaking proficiency interact with one another. Theoretically, it provides insight into how educators should implement AI tools in their language classrooms. As research, this adds to ongoing discussions on AI in education, though also calls for further investigation into the ethics and long-term effects of integrating AI into education at various scales.