Developing a Metric of Risk: Artificial Intelligence and Assessment

The rapid emergence of Generative Artificial Intelligence (GenAI) has created significant challenges for assessment integrity in Higher Education (HE). While institutions have initially relied on AI‑detection technologies, growing evidence demonstrates that these tools produce unreliable results and cannot serve as the primary safeguard for academic integrity. This paper responds to that gap by offering practical guidance for educators, assessors, course designers, and academic integrity specialists seeking to develop a metric of risk tailored to their own assessment contexts. Drawing on established risk‑assessment frameworks and contemporary research on AI in education, the paper proposes a dual‑dimensional model based on exposure and vulnerability, enabling users to evaluate how susceptible particular assessment types are to AI‑assisted misconduct. Extensions to the model, including behavioural, contextual, and temporal factors, are provided to support more nuanced analysis and future‑proofing. Through worked examples and design considerations, the paper demonstrates how the metric can inform proactive assessment redesign and promote authentic, ethical, and pedagogically aligned approaches in an AI‑enabled educational landscape. This work serves as a conceptual and practical resource for practitioners wishing to develop, adapt, or justify their own Metric of Risk in response to the evolving capabilities of GenAI.