AI-DRIVEN ADMISSION SYSTEMS AND THEIR IMPACT ON ADMISSION INTAKE AND INSTITUTIONAL DIVERSITY: A REVIEW

Abstract

With AI quickly making its way into higher education, many scholars are reevaluating how it decides the admissions. A systematic review of the peer-reviewed literature published during 2019 and 2025 throws light to the state of knowledge in relation to AI-led admission systems the details of operational architectures and documented impacts on volume of admission intakes and diversity of institution. Drawing on research published in Scopus and in Australian Business Deans Council (ABDC) classified journals, the review collates evidence on four linked themes (1) An adoption and functional architecture of AI admissions systems; (2) Predictive analytics and machine learning models for helping run the function of enrolment management; (3) Algorithmic bias with differential impact on racial, gender and socio-economic groups; and (4) Ethical and regulatory frameworks for using AI for admissions. The figures show that AI engines may increase throughput efficiency, reduce processing cost, increase predictive accuracy. But if historical data collects bias, then it may lead to reinforcement of same. Models that benchmark against historical enrolment norms focus on past limiting biases which aggravate the harm of Black, Hispanic, first-generation and low-income students. Academics studying Fairness, Accountability Transparency and Ethics (FATE) are constructing a normative vocabulary for reform. However, there’s a gap between the anticipations of policies and the technical implementation of these policies. A research agenda and recommendations for institutional policies to maximize alignment between equity-centred goals and AI-driven admission in higher education concludes our review.