As large language models (LLMs) and LLM-powered tools such as ChatGPT enter language classrooms, teachers face new opportunities and challenges in adapting their pedagogical practices. This study adopts a duoethnographic approach to examine how two university English as a foreign language teachers in China made sense of, negotiated and integrated LLM-powered tools over time. Guided by the cognitive-socio-technical framework, the analysis traces their experiences across cognitive, workflow and organisational levels through tandem writing and dialogic conversations. Findings reveal a four-stage trajectory, from initial excitement and early experimentation to critical reflection and balanced integration, showing how enthusiasm gradually gave way to reflective, teacher-led engagement. While LLM-powered tools enhanced efficiency and creativity, they also generated tensions related to academic integrity, teacher identity and institutional support. The study highlights the situated and developmental nature of teachers’ engagement with LLM-powered tools and underscores the importance of teacher agency, ethical awareness and institutional dialogue in supporting context-sensitive integration in English language education.
Kong et al. (Tue,) studied this question.