“All things are ready, if our mind be so”

Attitudes to STEM assessment redesign in the age of genAI

Abstract

In the two years since the launch of ChatGPT in November 2022, the higher education sector in Australia has experienced both the great potential of generative artificial intelligence (genAI) to transform learning and teaching, but also the serious threat to assessment security and assurance of learning that it poses. In light of recent regulatory calls for accountability in academic integrity, this article focuses on the issue of assessment security and redesign, given the affordances of readily available large language models. A nation-wide survey of STEM academics revealed that 37% of respondents had not even tested their assessments in a genAI app; that there was little consensus – and possible misunderstandings – concerning approaches for improving assessment security; and that sector recommendations were either not well understood, or not favoured, particularly those advocating a move towards program-level assurance of learning and only securing assessments at key points of a degree program. These results highlight the need for significant support and guidance for teaching academics—in terms of both creating a clear understanding of how best to respond to assessment redesign as future consensus emerges, and institutional workload in implementing those responses —in order for challenges posed by genAI to assurance of learning to be met successfully.