From Principles to Practice: An Operational Framework for Equitable Artificial Intelligence in Educational Assessment

The integration of artificial intelligence (AI) into educational assessment presents both transformative opportunities and critical ethical challenges. While AI promises enhanced personalisation and efficiency, these systems can perpetuate and amplify existing educational inequities through algorithmic bias. This conceptual paper argues that effective governance requires moving beyond a static checklist of principles toward a dynamic, problem-driven cycle. We propose a novel ethical framework structured around three sequential questions institutions must address: a goal-setting question (Fairness), a knowledge question (Transparency), and a governance question (Accountability). We then operationalise this framework through a lifecycle model, demonstrating how these ethical imperatives can be embedded at the pre-processing, in-processing, and post-processing stages. Adapting established principles from educational measurement and embedding them within concrete governance structures, the paper provides an actionable pathway for institutions to ensure AI assessment systems serve as tools for educational equity rather than instruments of disadvantage.