‘A bit of chaos and madness’: The AI Assessment Scale and the work of assessment reform

As generative AI (GenAI) disrupts traditional assessment, universities are moving away from simple prohibition toward structured frameworks. This study investigates the implementation of the AI Assessment Scale (AIAS)—a tool designed to help educators integrate AI into their assessments while maintaining academic integrity. The researchers conducted five focus groups with 30 academic staff across a private international university in Vietnam and a public university in the UK. Using hybrid thematic analysis and the lens of Critical AI Literacy, the study explores how faculty experience these frameworks in practice.

Faculty generally valued the AIAS as a practical tool that provided a shared vocabulary for discussing AI. It helped legitimize the use of GenAI in the classroom, moving the conversation beyond 'cheating' and toward pedagogical integration. By providing clear boundaries, the scale offered staff the confidence to experiment with AI tools without fear of reprimand. However, the study notes that for some, the scale was seen as a secondary influence; many staff felt their pedagogical shifts were driven by the rapid evolution of GenAI itself rather than the framework alone.

Despite the benefits, the study highlights significant challenges in moving from policy to practice. Implementation was often hampered by:

• Policy Gaps: When the AIAS was introduced without clear mechanisms for evaluation or monitoring, it created a sense of 'chaos and madness' for staff.
• Compliance vs. Pedagogy: There is a risk that the AIAS becomes a mere 'compliance layer'—a box-ticking exercise used by institutions to demonstrate they have 'done something'—rather than a tool for meaningful assessment redesign.
• Workload and Capacity: Staff reported that without adequate training and peer-to-peer sharing, the burden of managing AI integration remained high, leading to uncertainty about how to maintain fairness and academic standards.