Transforming Scholarship Selection: A Review of AI-Driven Matching Algorithms and Transformer Models

Selecting a candidate for a scholarship requires careful evaluation that can easily promote or restrict the equity and accessibility of education. However, traditional selection methods often rely on subjective evaluations, manual processes, and limited scalability, resulting in inefficiencies and potential biases. As the volume of scholarship applications increases, artificial intelligence (AI)-driven solutions utilizing deep learning and transformer models have emerged to enhance the selection accuracy, fairness, and efficiency. This work-in-progress conducted a review of existing literature relevant to the topic using the PRISMA guidelines. It investigates the additional advantages of each algorithm and their ethical implications, including bias reduction, privacy concerns, and transparency. To ensure reliable AI implementation in scholarship selection, the challenges related to algorithmic fairness, explainability, and compliance with ethical requirements must be addressed. The findings indicate that AI-driven approaches substantially improve selection accuracy, streamline application processing, and reduce the human workload. For future research, human-AI collaboration in AI-based scholarship selection is proposed to enhance fairness, transparency, and accountability in AI-powered decisions. Therefore, refining explainable AI models and integrating fairness-aware mechanisms should be undertaken to enhance trust in automated decision-making systems.