Low-Proficiency ESL Learners’ Emotional, Cognitive, and Metacognitive Engagement in Face-to-Face Versus AI-Mediated Spoken English Communications

Abstract

This multimodal phenomenological study investigates the lived experiences of low-proficiency ESL undergraduates’ emotional, cognitive, and metacognitive responses during human versus AI chatbot-partnered spoken interactions using the target language of the learners, i.e., English. A multimodal visual analysis approach, grounded in thematic coding, supported the analysis of participant drawings, and the data from the semi-structured focus group discussions were thematically analyzed. The findings reveal a significantly positive response to including voice-enabled AI-chatbots as potential L2 speaking partners due to their non-judgmental, uncritical, accommodating, and facilitative potential, drastically reducing Foreign Language Anxiety (FLA), cognitive overload, fear of making mistakes, being judged, and ridiculed, all of which were present during their spoken interactions in English with a proficient speaker. Situated in a postcolonial context, the implications of reticence in speaking English yield significant findings that highlight nuanced, problematic, and ideologically and socially laden implications as they unravel the internalized power asymmetries and linguistic marginalization rooted in unwarranted colonial legacies. The study’s major contribution lies in its depiction of the underrepresented postcolonial perspectives of linguistically disempowered individuals during their L2 speaking skills development. While reflecting on the potential of AI as an enabling resource for supporting the development of spoken English among beginner-level adult ESL learners, this study also draws attention to the structural inequalities within a postcolonial community that should be addressed with the understanding that AI alone cannot mitigate reticence in English speaking.