Conceptualizing critical AI literacy in writing education: power dynamics in Chinese EAL students' negotiations with GenAI

Generative artificial intelligence (GenAI) is rapidly reshaping writing education, yet discussions of AI literacy often foreground tool knowledge and strategic use without adequately theorizing how power operates in AI-mediated literacy practices. This article conceptualizes Critical Artificial Intelligence Literacy (CAIL) in writing education by integrating power dynamics into existing accounts of digital, critical, and AI literacies. Drawing on a qualitative case study of five Chinese English as an Additional Language (EAL) students, the study analyzes screen recordings, written drafts, and semi-structured interviews to examine how students negotiate AI across algorithmic outputs, linguistic expression, cultural reference, and learner agency. Findings show that AI's authoritative-sounding language and uneven reliability position students as vigilant evaluators who corroborate information, manage citation risk, and recalibrate trust. Students also navigate tensions between AI-driven standardization and personal voice, and between homogenized cultural framings and situated cultural meaning-making. Across these negotiations, writing emerges as a posthuman, distributed activity in which agency is co-produced within human-AI assemblages, making empowerment contingent on learners' capacity to critically attribute, contest, and strategically incorporate AI contributions. The article argues that CAIL should be conceptualized as the development of AI-related knowledge and strategies together with reflective understanding of the power dynamics shaping AI-mediated writing.

Duke Scholars

Author Yachao Sun DKU Faculty