INTEGRATING SELF-MONITORING METACOGNITIVE TOOLS WITH VOICE AI FOR REDUCED SPEECH HESITATION IN EFL CLASSROOMS

Abstract

Speech hesitation hinders spontaneous oral communication among EFL learners. While Conversational Voice AI provides interactive practice, standalone use lacks mechanisms to foster self-regulated speech production. This study evaluated the efficacy of integrating self-monitoring metacognitive tools (visual pause analytics, reflection logs, and goal sheets) with Voice AI to reduce speech hesitation among intermediate EFL university students. Utilizing an 8-week mixed-methods design, participants were divided into experimental (Voice AI + metacognitive tools) and control (Voice AI alone) groups. Pre- and post-intervention monologues were analyzed acoustically via Praat software for filled pauses, unfilled long pauses, and speech rate, alongside qualitative thematic analysis of reflection logs. Results demonstrated a significant reduction in filled and unfilled pauses and a substantial increase in speech rate in the experimental group compared to the control group. Qualitative findings confirmed that visual feedback heightened temporal self-awareness, shifting learners from silent freezing to active paraphrasing strategies. Integrating metacognitive tools with Voice AI transforms passive technology use into active self-regulation, offering a scalable framework for enhancing EFL oral fluency