Student authorship and linguistic diversity in English education with generative artificial intelligence

Abstract

This theoretical essay examines the pedagogical and analytical conditions for understanding student authorship in written and multimodal English productions mediated by generative artificial intelligence. Its argument builds on a master's dissertation on digital technologies in lower secondary education and extends questions formulated in a doctoral research proposal focused on Brazilian public upper secondary education. An interpretive bibliographic approach brings together purposefully selected studies on multiliteracies, translanguaging, critical artificial intelligence literacies, and language education, alongside the dissertation's educational proposals. The essay argues that the formal quality of a final product is insufficient to establish authorship or learning, since it may conceal decisions delegated to a system and the erasure of local repertoires. Its conceptual contribution is the authorial decision episode as an analytical unit connecting an initial formulation, an automated suggestion, a student's intervention, and its situated justification. This proposal is developed into criteria addressing linguistic diversity, the verifiability of claims, and responsibility for textual circulation. Hypothetical examples and guidelines for process portfolios and formative assessment illustrate the argument without claiming empirical effectiveness. The essay concludes that critical integration of artificial intelligence requires institutional access and teacher mediation that enable students to question and transform available resources rather than merely reproduce polished outputs.