Abstract:

This paper examines how assessment in higher education is changing in response to generative AI, using Bifröst University in Iceland as a case study. As a university built on a digital learning ecosystem with extensive online teaching, Bifröst provides a relevant context for exploring how assessment methods can be adapted to maintain academic rigour and integrity. The objective of the paper is to identify which assessment approaches remain effective in AI-rich learning environments and to evaluate the potential of more diverse and reflective assessment formats.

The study is based on an action research approach that draws on the experience of a university professor who has used a wide range of assessment methods before and after the emergence of AI tools. The analysis compares the number and type of assessment methods used across these two periods, with particular attention to reflective formats such as podcasts and learning diaries.

The findings indicate a clear shift towards greater diversity in assessment after the arrival of AI. Approximately ten assessment methods were used before AI and a similar number after AI, but only four methods overlap between the two periods. This suggests substantial experimentation and redesign in recent years. The results further indicate that podcasts can support argumentation, peer-to-peer learning, and communication skills relevant to professional contexts, while diaries remain valuable for individual reflection and for documenting students’ learning processes in a more personal and authentic way.

The paper concludes that higher education institutions need a broader and more deliberate mix of assessment methods in response to AI. The Icelandic experience suggests that both revised traditional tools and newer reflective formats can support deeper learning, practical application, and more robust forms of assessment in digitally mediated higher education.

Keywords:

Higher education, assessment, generative AI, podcasts, learning diaries, online learning