Authorship in the Age of AI: Reclaiming Voice and Process in the EFL Writing Classroom

This paper explores the pedagogical challenges and opportunities presented by generative artificial intelligence (AI) in the English as foreign language (EFL) writing classroom. As large language models (LLM) become increasingly integrated into student workflow, traditional product-based assessments lose their reliability as indicators of individual learning. This paper argues for the strategic rethinking of writing instruction, incorporating AI while preserving the cognitive and ethical dimensions of authorship. It begins by outlining how product-based, process-based, and genre-based writing approaches carry varying degrees of vulnerability to AI interference. Drawing from empirical research and classroom practice, it then illustrates how AI may be a detriment to both learning and writing. In response, the TRACE Model is introduced: a five-phase instructional framework that guides users through drafting, feedback, and reflection while regulating when and how AI is used. In this model, AI is positioned as a late-stage tool for reflection and feedback—not initial composition—so that student choice, teacher scaffolding, and structured reflection help maintain engagement, build metacognitive awareness, and reinforce the human core of the writing process.