Exploring patterns of AI-powered machine translation use in second language writing: a translanguaging perspective

Machine translation (MT) has emerged as an artificial intelligence (AI) technology in second language (L2) writing. Research to date has predominantly examined students’ MT use in L2 writing in quasi-experimental settings, with limited focus on its use in naturalistic contexts. This qualitative study, informed by translanguaging theory, investigated the patterns of MT use among students and the factors shaping these patterns in authentic settings. The study was conducted at a Chinese university with undergraduate participants from both English and non-English majors. Data were collected in two stages. In Stage One, the participants completed a narrative frame, with interviews conducted for a portion of them. Stage Two focused on observing actual MT practices, where the students shared screen recordings of their MT use, kept reflective journals, and engaged in individual interviews. Qualitative analysis was then conducted. The findings reveal an extensive use of MT in L2 writing, with four major patterns of MT use: using MT as a language bridge, an alarm bell, a critical sword, and a creativity springboard. Three ideological factors—monolingual ideology, standard language ideology, and a translanguaging stance—were found to shape these patterns of MT use. Based on the findings, this study proposes an ideological conceptual framework of students’ MT use to illustrate the interactions between patterns and ideologies. This framework provides valuable insights for integrating AI-powered MT as a pedagogical tool in L2 classrooms and suggests different pathways for language educators to support students’ effective use of MT in L2 writing.