Voices in feedback: A dual-theoretical exploration of learners’ perception of teachers versus AI-powered tools

Abstract

Feedback plays a central role in pronunciation learning by helping learners notice and adjust their phonological output. With the growing use of artificial intelligence (AI) and automatic speech recognition technologies, learners increasingly receive feedback from both human and automated sources. This qualitative study explores how five Vietnamese English-major undergraduates perceived pronunciation feedback from AI-powered tools and teachers during a seven-week learning cycle, and how they integrated these feedback sources in their practice. Drawing on Interactionist Second Language Acquisition theory and the Feedback Literacy framework, data were collected from learning diaries, written reflections, and semi-structured interviews. The findings show that learners valued AI-power tools for their immediacy, accessibility, detailed pronunciation feedback, and low-pressure practice environment. However, they also questioned AI feedback when corrections appeared excessive or unclear. In contrast, teacher feedback was perceived as more reliable, explanatory, and personally relevant, although public correction sometimes caused anxiety. Learners tended to use these tools for initial rehearsal and repeated practice, while relying on teachers for clarification, validation, and deeper understanding. The study suggests that learners strategically combined AI and teacher feedback to balance efficiency, interpretation, and emotional support in pronunciation learning.