Artificial intelligence in higher education assessment as an ecosystem of pedagogy governance and legitimacy

Abstract

Artificial intelligence (AI) is rapidly reshaping higher education assessment through applications such as automated feedback, adaptive testing, learning analytics, automated scoring, and generative support. Yet scholarship on AI-enabled assessment remains fragmented across technical, pedagogical, ethical, governance, and professional perspectives, limiting the field’s capacity to explain how these dimensions interact in practice and how misalignment among them affects educational legitimacy. This paper develops the AI-Assessment Ecosystem Model, a conceptual ecosystem framework for analysing AI in higher education assessment as a sociotechnical and institutional phenomenon rather than a discrete technical intervention. Drawing on a conceptual synthesis of literature published primarily between 2019 and 2024, and informed by foundational work on assessment validity, feedback, institutional legitimacy, organisational change, inclusive design, and computer-supported collaborative learning, the paper identifies five interdependent domains that shape AI-enabled assessment: technological innovation, pedagogical alignment and learner needs, institutional strategy and governance, faculty and professional readiness, and ethical and inclusive implementation. The paper theorises legitimacy as a relational and contested process through which AI-supported assessment must be justified to students, faculty, institutions, regulators, and other stakeholders as educationally valid, professionally defensible, ethically acceptable, inclusive, and organisationally accountable. The framework argues that AI in assessment cannot be adequately evaluated by technical capability, implementation success, or ethical compliance alone; it must be understood through the relations, tensions, and potential misalignments among educational purpose, organisational conditions, professional practice, student agency, and technological design. The model is offered as a conceptual ecosystem framework for guiding future research, critique, and institutional decision-making.