Whose feedback matters? Exploring human and AI-supported writing feedback practices in a South African writing centre

Abstract

In the context of the growing reliance on artificial intelligence (AI) tools for academic writing, this article explores how undergraduate students at a South African writing centre perceive and use feedback from both human tutors and AI-based tools. Drawing on Feedback Literacy and Critical Digital Pedagogy, the study investigates the strengths and limitations of AI-generated and human feedback, as well as how resource constraints shape student practices. Using qualitative interviews with 10 undergraduate students across diverse disciplines, the findings reveal that while AI tools offer efficiency for surface-level corrections, students express a strong preference for the relational, dialogic, and context-sensitive nature of human feedback. Students strategically blend AI and tutor feedback, demonstrating emerging feedback literacy, yet their practices remain constrained by infrastructural barriers and algorithmic biases. The study highlights the risk that AI-generated feedback may reinforce educational inequalities if digital divides and linguistic diversity are not adequately addressed. It calls for feedback ecologies that balance technological innovation with human care, supporting equitable, and inclusive writing development in under-resourced, multilingual higher education settings.