The Impact of Ai-Driven Personalised Learning on Enhancing Student Academic Performance

Abstract

Artificial intelligence (AI)-driven personalised learning has altered the design of educational environments by enabling adaptive content selection, automated feedback, learner profiling, progress monitoring, and personalised learning pathways. This study develops and tests a conceptual framework linking AI-driven personalised learning with student academic performance through perceived content relevance, perceived feedback utility, self-regulated learning autonomy, and AI-driven learning engagement. The framework is grounded in the Technology Acceptance Model, self-regulated learning theory, and research on adaptive learning. A quantitative cross-sectional design is specified for higher-education students using a structured five-point Likert questionnaire. The proposed empirical procedure includes reliability analysis, convergent and discriminant validity assessment, descriptive statistics, correlation analysis, and multivariate regression or partial least squares structural equation modelling. For manuscript development, an illustrative dataset of 300 cases is used to demonstrate the analytical structure. The illustrative results produce a model R² of 0.639 for student academic performance, with positive standardized coefficients for perceived content relevance (β=0.176), perceived feedback utility (β=0.199), self-regulated learning autonomy (β=0.245), and AI-driven learning engagement (β=0.218), all with p<0.001 in the simulated regression. The descriptive pattern places AI-driven learning engagement at a mean of 3.75, followed by perceived content relevance at 3.72, self-regulated learning autonomy at 3.68, and perceived feedback utility at 3.61 on a five-point scale. The findings section is explicitly illustrative rather than empirical. The paper establishes a research-ready structure for examining whether and through which learning mechanisms AI personalisation is associated with academic performance. The framework also recognises that AI use does not automatically constitute effective learning: the educational value of AI depends on the quality of personalisation, feedback, student agency, assessment design, and the extent to which AI supports rather than substitutes for cognitive activity.