Beyond Task Delegation: Reconceptualising Instructional Design to Measure Competence in the Age of Generative AI

Abstract

The rapid proliferation of Large Language Models (LLMs) has invalidated traditional content-recall assessments, acting as a catalyst for a systemic transition towards Competence-Based Education (CBE). This paper contends that the role of faculty must transcend content dissemination, focusing instead on observable human competencies associated with Metacognition and Self-Regulated Learning (SRL). Drawing upon Behaviourism and Bloom's Taxonomy, we propose the Reverse-Flipped Model. This conceptual framework, supported by illustrative disciplinary cases in History, Computer Science, Digital Photography, and Composition, utilises generative tools for low-stakes production to repurpose pedagogical time for high-stakes human intervention. Prompt Literacy is conceptualised as an observable instructional competency that externalises planning, monitoring, evaluation, and strategic intervention behaviours during human-AI collaboration. Implementation necessitates the use of Analytic Rubrics and Automated Process Portfolios (APP) to document observable demonstrations of Prompt Literacy and strategic engagement over final product quality. The framework addresses institutional sustainability and the moral imperative of digital equity. The primary contribution of this framework is a structural shift that enhances assessment validity by ensuring student success is defined by verifiable mastery of human-led strategic intervention rather than machine-generated output.