Beyond essay primacy: written assessment regimes, generative AI, and evaluative visibility in higher education

This article examines how essay-, report-, and dissertation-based assessment regimes shape what becomes visible as student learning in UK higher education. It focuses on generative artificial intelligence not as a reason to abandon written assessment, but as a prompt to reconsider what written work can reasonably be taken to evidence. The article introduces the concept of evaluative visibility to describe how assessment makes some forms of learning, judgement, and capability available for recognition and reward, while leaving others less directly visible. Drawing on a qualitative documentary analysis of UK policy texts, assessment artefacts, and guidance on generative AI, the article shows that written assessment remains valuable for evidencing disciplinary knowledge, argument, synthesis, and use of evidence. However, where the finished written product is expected to stand for a wider range of capabilities, including contextual problem-framing, audience responsiveness, coordination, and accountable judgement, the evidential basis for such claims may be less secure. The article contributes to debates on assessment validity by shifting attention beyond authorship and academic integrity towards the broader question of what assessment makes visible, and what may need to be evidenced in other ways, in the generative AI era.