Preserving epistemic authenticity: process-oriented assessment in the age of generative AI

Counts differ because each database indexes a different set of publications. We treat OpenAlex as the canonical count; Google Scholar is not shown (no API, and crawling it violates its ToS).

The rapid advancement of generative artificial intelligence (AI) has introduced significant challenges for higher education (HE) assessment practices, particularly those relying on unsupervised written tasks. As AI systems become increasingly capable of producing sophisticated academic text, traditional product-oriented assessments face growing pressures to reliably capture ‘authentic’ student learning, reasoning and disciplinary engagement. In response, we argue for a shift towards more process-oriented assessment approaches that foreground how learners interpret, justify and construct understanding over time. Drawing on contemporary scholarship surrounding generative AI, authentic assessment, evaluative judgement and assessment redesign, we examine how process-oriented approaches can offer a more intellectually rigorous and sustainable framework for assessment and evaluation. We also consider broader pedagogical and institutional implications associated with this shift, including educational equity, emotional labour, workload sustainability and what we describe as epistemic authenticity. Additionally, we outline key principles for designing such approaches, including reflective justification, staged task design, dialogic engagement and evaluative transparency. Finally, we conclude that reorienting assessment towards the evaluation of reasoning, interpretation and knowledge construction may provide a more sustainable, intellectually meaningful and pedagogically robust response to the challenges posed by generative AI within HE.