Enhancing EFL Learners’ Competence in Translating Chinese Multiword Expressions: A Process-Oriented Instructional Model for the AI Era

As generative artificial intelligence rises and reshapes the way learners deal with translation tasks, there is a growing need for new teaching models that highlight human-AI collaboration. In response to this demand, the study proposes and verifies a process-oriented teaching model, which aims to cultivate the strategic ability of EFL learners to translate Chinese multiword expressions. A quasi-experimental design was adopted. Two intact undergraduate classes of translation major were selected and randomly assigned to the experimental group and the control group, with 32 students in each group. Data were collected through screen recordings, stimulated recall interviews, translation quality assessments and in-depth interviews. Quantitative analysis showed that the workflow efficiency and translation quality of the experimental group were significantly higher than those of the control group. Qualitative data further indicated that the students in the experimental group learned to guide AI more strategically and established their professional identity as strategic coordinators rather than passive tool users. This study provides a practical framework for integrating human-AI collaboration into the cultivation of translation talents, and further enriches the theoretical connotation of translation competence in the era of AI.