Assessing multimodal, multilingual and GenAI academic literacy

Abstract

This article critically reflects on the assessment outcomes of a semester-long academic literacy module designed for first-year Humanities students at the University of Pretoria. It presents both qualitative and quantitative findings from a pedagogical approach that integrated GenAI-mediated visual knowledge construction with conventional academic literacy practices, particularly academic writing. The data derive from the AI-Guided Visual Literacies: Illuminating Disciplines for Conceptual Understanding project, which invited students to incorporate visual elements in articulating their pre-university and emerging disciplinary identities. Findings highlight both the affordances and limitations of this multimodal approach. On the one hand, the integration of visual modalities enabled students to transcend the constraints typically associated with unimodal textual expression. On the other hand, a 15.5% performance decline was observed when students transitioned to more traditional, exclusively written forms of academic discourse, such as the literature review. The findings reveal that while multimodality holds promise for enhancing student engagement and self-representation, conversely, its integration into academic literacy curricula demands a nuanced appreciation of students’ varying degrees of readiness for university-level knowledge production. The article concludes by proposing pedagogical strategies that bridge conventional academic genres with multimodal and multilingual forms of meaning-making, including GenAI.