A Threat Assessment Framework for Screening the Integrity of University Assessments in the Era of Large Language Models

Since late 2022, the sudden growth in the availability and capabilities of generative artificial intelligence tools, such as Large Language Models, has raised concerns about the threat they pose to the integrity of assessment in educational institutions. Such models are constantly evolving and improving, making the task of understanding exactly what they can do more difficult. Recognising this challenge, this paper establishes a Large Language Model exposure framework to qualitatively and quantitatively examine the assessment strategies of university modules to provide a high-level estimated indication of the exposure of these modules to potential dishonest use of such models in the completion of their assessments and coursework. This framework may be used and adapted when planning and reviewing teaching and learning practices and policies.