Using generative artificial intelligence to reimagine feedback in higher education: a collaborative autoethnography

Abstract

This article examines the potential of generative artificial intelligence (GenAI) to reshape feedback practices in higher education. It reports on a proof-of-concept study in which teacher educators recorded oral feedback. These recordings were uploaded to a customised ChatGPT model and restructured in alignment with the assessment rubric. Educators then edited the text to ensure accuracy, appropriate tone, and alignment with the rubric. Using collaborative autoethnography, we analysed our experiences to explore how GenAI reshaped the feedback process. Initial implementation increased workload due to technical challenges and substantial editing demands. Over time, efficiencies improved as prompts and workflows were refined. Some educators reported reduced cognitive demands when speaking rather than typing, although labour shifted towards verification, tone calibration, and rubric alignment. While GenAI output imposed a coherent structure on spoken commentary, human oversight remained essential. Ethical considerations relating to authorship, transparency, and professional responsibility were central throughout. We argue that meaningful integration of GenAI into feedback practices requires careful design, sustained human oversight, and explicit ethical reflection. This study raises important questions about assessment authorship, professional identity, and evolving assessment processes in higher education.