Exploring Low-Proficient ESL Undergraduates’ Responses to Face-to-Face and AI-Mediated English-Speaking Experiences

Abstract

The omnipresence of educational technologies has become an undeniable reality in the post-COVID era. Generative Pre-trained Transformer (GPT) technology in formal and informal teaching and learning contexts has resulted in a pervasive and often controversial presence in the contemporary education landscape, driven largely by the rapid popularization of OpenAI, specifically ChatGPT. This phenomenological study intends to capture and reflect on the lived experiences and emotional realities of low-proficient adult ESL learners when communicating in the second language, i.e., English, as it compares their experiences during face-to-face spoken interactions and those of ChatGPT as a communicative partner. Through purposive sampling, 84 visual arts undergraduates were selected for the study. Using the multimodal approach, the arts-based visual data were gathered from students depicting their experiences during conversations with English speakers and ones where they speak directly to ChatGPT’s AI assistant. The narrative exploration was done by collecting data via group discussions, whereby participants reflected on and articulated their feelings during both human and AI-mediated communication. In addition, visual data were gathered from the sample, whereby they shared drawings of their visualization of the experiences during face-to-face vs AI-assisted spoken communication. Major study findings uncover the stark contrast between the participants’ experiences with the two modes of communication with their artwork and verbal exchanges reflecting powerful and recurring themes of the fear and anxiety present during face-to face conversations indicating such instances as high-stakes contexts with intense pressure due to the power dynamics placing the L2 learner in a vulnerable and disadvantaged position. However, the AI-mediated interactions appear to create different emotional and learning spaces for the students, which are in stark contrast to the anxiety-ridden imagery associated with their face-to face spoken interactions. Given the obvious reticence to English speaking among L2 learners, it is a timely requirement to engage in further research focusing on AI-assisted language support and its transformational potential.