A Prescriptive Learning Analytics Framework: Beyond Predictive Modelling and onto Explainable AI with Prescriptive Analytics and ChatGPT

Prediction alone is insufficient for helping struggling students: The paper highlights that most learning analytics research focuses only on predicting which students will drop out, but neglects the crucial step of explaining why these predictions are made or what to do about it—a gap that limits the practical value of these systems for educators and students.

Explainability is essential for stakeholder trust: The framework emphasizes that machine learning models used in education must be transparent and interpretable, allowing teachers, advisors, and students to understand the reasoning behind at-risk predictions rather than treating algorithms as "black boxes," which is critical for responsible deployment in educational settings.

AI should move from diagnosis to actionable guidance: Beyond identifying at-risk learners, the paper advocates for "prescriptive analytics" that generates specific, evidence-based recommendations for intervention—such as tailored advice on study strategies or resource access—making AI systems more directly useful for improving student outcomes.

Large language models can enhance human communication of AI insights: The integration of ChatGPT-like technology demonstrates how AI can translate complex algorithmic outputs into clear, personalized, human-readable feedback that stakeholders can actually understand and act upon, bridging the gap between technical AI systems and real-world educational practice.

Responsible AI in education requires both transparency and actionability: The framework represents a shift toward more ethically sound educational AI that doesn't just make decisions about students but explains those decisions and empowers educators with concrete next steps.