THE ROLE OF AI IN ENHANCING EFL LEARNERS' SPEAKING SKILLS

Abstract

A persistent gap remains between digital speaking tools and the way they actually reshape oral performance in EFL classrooms. This paper examines how artificial intelligence supports speaking development through qualitative content analysis and comparative reading of classroom reports and pedagogical descriptions. The material was grouped into three recurring feedback modes: immediate pronunciation correction, guided dialogic prompting, and adaptive performance tracking. Across the reviewed cases, the strongest gains appeared in fluency stabilization, clearer segmental pronunciation, and reduced hesitation during short speaking tasks, while weak systems mainly simulated interaction without sustained learner uptake. The analysis also showed that feedback combining timing, specificity, and follow-up prompts produced four stable improvement effects: self-monitoring, reformulation, confidence growth, and longer turns. The study concludes that AI contributes most when it functions as an interactive scaffold rather than a replacement for human-mediated speaking practice.