AI-Based Corrective Feedback in EFL Interactive Speaking: Insights from Interactionist SLA Theory

Abstract

This study explores the effectiveness of artificial intelligence (AI)-driven corrective feedback in enhancing the speaking accuracy and fluency of learners of English as a Foreign Language (EFL), drawing upon principles from Interactionist Second Language Acquisition (SLA) theory. The investigation seeks to understand the impact of AI technologies on learners within language education settings, particularly during oral task performance. The research involved 20 EFL students who engaged with varying levels of AI language tools. Data collection was conducted through semi-structured interviews and group discussions. A thematic analysis was performed using NVivo software. Findings indicate that participants showed notable improvements in grammatical accuracy and spoken fluency as a result of engaging with AI-based training. However, students identified limitations in the AI feedback, particularly the lack of tailored explanations and affective support, which are typically offered by human instructors. Although AI demonstrated considerable capability in delivering corrective input, learners continued to prefer teacher-led evaluations. This study contributes to the academic discourse in the EFL domain by critically examining both the advantages and shortcomings of incorporating AI into speaking instruction. It suggests that blending conventional pedagogical methods with AI tools may foster more effective spoken language development.