Hongqin Li

Lin Pan

Beijing Normal University, China

Abstract

This study investigates how English major students collaborate with generative artificial intelligence (AI) chatbots in English as a Foreign Language (EFL) writing assignments. Employing Actor-Network Theory (ANT) and critical discourse analysis, this research focuses on a symmetrical treatment of both human and nonhuman aspects of students’ diverging digital literacy practices. Data were collected from 12 participants through writing drafts, ChatGPT chat logs, screen recordings, and stimulated recall interviews at a university in China. In producing students’ AI-assisted writing networks, the analysis illustrated how the authenticity of ChatGPT’s output was established, stabilized and contested with students’ shifting understandings towards an idealized native-like resemblance in AI’s language norms and its ‘illusionary’ objective stance. As students continuously negotiate authority between native-like authenticity and their original voices, this study highlights how their strategic decision-making in this process is conflated with academic and occupational pressure towards competitive advantages. The study thus reveals how students’ engagement with ChatGPT is not merely a technical process but also a complex network of language ideologies, academic expectations, and personal identities. The need to cultivate critical AI literacy among students is discussed, and this study also calls for reconsidering EFL writing instructions and assessment criteria in the era of generative AI.

Keywords

Artificial intelligence, English writing, actor network theory, critical discourse analysis