From Crisis to Opportunity: Reimagining Academic Assessment for the AI Era

The integration of generative artificial intelligence (AI) into higher education presents a critical juncture for assessment practices. Traditional methods, such as take-home essays and unsupervised coursework, are increasingly vulnerable to AI-generated submissions, challenging their validity and raising profound questions about academic integrity (Cotton et al., 2023). This paper argues that the prevailing response focusing on AI detection and restriction is unsustainable. Instead, we must seize this moment as a catalyst to fundamentally redesign assessments for authentic, humancentric learning. Specifically, this paper proposes an AI-resilient assessment design framework that aligns assessment tasks with cognitive, reflective and ethical capabilities that remain distinctly human. Moving beyond a deficit model, this paper explores a framework for "AI-resilient" assessment that prioritises the cognitive processes education seeks to develop. Drawing on principles of authentic assessment (Boud & Falchikov, 2007) and contemporary digital literacy frameworks (JISC, 2023), we will present and critique a spectrum of alternative approaches. These include scaffolded project work that evidence iterative development, reflective portfolios that articulate metacognitive journeys, and in-person vivas that probe critical reasoning. Critically, the paper contends that successful adaptation requires enhancing assessment literacy for both staff and students. This involves co-creating transparent guidelines on ethical AI use and fostering a shared understanding of how AI can function as a collaborative tool rather than a substitute for learning. Ultimately, the paper proposes that the AI era requires a necessary progression towards assessments that measure distinct human skills such as critical thinking, ethical reasoning, and creative collaboration within learning spaces that are essential for graduate success in a digitally transformed world (OECD, 2021).