Exploring EMI teachers’ agency in addressing language-related challenges with ChatGPT: a multiple case study
Wenyun Jia, Jack Pun
Abstract
Teachers have begun to use generative artificial intelligence (GenAI) chatbots, such as ChatGPT (hereafter, GPT), as pedagogical tools to help learners acquire English language skills. In English-medium instruction (EMI) classrooms, content teachers may lack sufficient and contextualized methods to address students’ language challenges, which could impact both content and language learning outcomes. However, research on how EMI teachers use GenAI chatbots to accommodate language learning needs is limited. To fill this gap, the study adopts an ecological perspective of teacher agency to explore how three EMI content teachers exercised agency in using GPT to address language-related teaching challenges in their classrooms. The study further examines the factors influencing their agency. The data primarily consist of teachers’ GPT interaction logs and responses from interviews regarding their experiences and perceptions of GPT. The findings reveal that the teachers exercised their agency by 1) addressing perceived language-related teaching challenges, 2) leveraging GPT affordances, and 3) mitigating school-level and GPT-level constraints. This study highlights the diverse and dynamic pathways of agency enactment, especially how they are influenced by teachers’ past technology experiences, social contexts, evaluation of GPT affordances and constraints, and aspirations regarding students’ success and personal growth. Notable, teachers’ initial engagement with GPT and their ongoing evaluation in response to contextual constraints play a crucial and ongoing role. These insights can inform efforts to facilitate productive GPT–teacher collaborations in addressing language teaching issues in EMI contexts.