Fair Marking in the Generative AI Era: Introducing the Master’s Dissertation Marking Framework

This paper presents the Master’s Dissertation Marking Framework (MDMF), a longitudinally developed framework designed to support fairer and more transparent master’s dissertation assessment. The framework was developed through a multi-phase, design-based research framework, comprising a literature review, a survey and in-depth interviews (2022) conducted prior to the emergence of generative AI, and follow-up empirical phases between 2023 and 2025. Across these phases, the framework evolves from an initial focus on procedural consistency and bias mitigation to a broader sociotechnical perspective that incorporates ethical boundaries, professional judgement, institutional responsibility, and the disruptive effects of generative AI on assessment practice. The paper traces the progression of the framework to MDMF Version 5, the final iteration, which consolidates six interdependent components: ethical boundaries and AI policy clarity; fairness and equity issues; pre-marking tasks and calibration; marker allocation; marking processes, culture, and well-being; and technology as both enabler and disruptor. Drawing on empirical evidence from academic staff involved in MSc dissertation marking in the post-generative-AI context, the framework brings together these components to address both longstanding and emerging challenges in assessment. The findings demonstrate that fairness in dissertation marking cannot be achieved through procedural mechanisms or technological solutions alone. Instead, the MDMF supports fairer assessment by structuring human judgement, enabling calibration, and clarifying ethical boundaries in AI-mediated contexts. The framework offers a coherent yet adaptable model for institutions seeking to maintain valid and defensible assessment practices in the age of generative AI.