Generative AI in Second Language Video Composing: Investigating Students' Critical Digital Literacies and Multimodal Redesign

Generative artificial intelligence (GenAI) in digital multimodal composing (DMC) presents both opportunities and challenges for second language (L2) research and pedagogy. While GenAI tools provide linguistic and multimodal support, they also raise concerns about biased information, culturally generic outputs, and technological overreliance. This qualitative study investigates L2 students' enactment of critical digital literacies (CDL) during GenAI‐assisted video composing. Drawing on CDL as a theoretical orientation and the analytical heuristic of critical awareness, critical identities, critical relationships, and critical actions, the study examines how students negotiated and revised AI‐generated video representations. Using multimodal critical discourse analysis of screen recordings, final video artifacts, focus group interviews, and reflective essays from 20 L2 students, the findings reveal four interrelated forms of CDL enactment. Students demonstrated critical awareness of GenAI's culturally limited outputs, authored critical identities through multimodal redesign, negotiated critical relationships with peers and GenAI, and engaged in localized critical action against cultural erasure. However, students' varied responses also highlight the uneven opportunities for CDL enactment across student groups. These findings offer theoretical and pedagogical implications for supporting L2 learners' critical engagement with algorithmic assumptions, cultural biases, and multimodal representations in GenAI‐assisted DMC.