The Influence of AI-Facilitated Personalised Education on Improving Student Academic Performance

Abstract

AI has revolutionised the landscape of personalised learning, empowering learning systems to analyse learner data, tailor learning paths, provide feedback, suggest learning materials, and pinpoint learning challenges. This paper aims at answering the main question: What is the association between AI-facilitated personalisation and students' academic achievement, and what are the implementation conditions that influence this association? This paper introduces a systematic, evidence-based review that integrates findings from peer-reviewed systematic reviews, scoping reviews, meta-analyses, and selected empirical research in the fields of artificial intelligence in education, adaptive learning, intelligent tutoring, personalised recommendations, and digitally-delivered feedback. The findings suggest that personalised adaptive learning is linked with higher academic achievement in a significant number of the studies examined. A 2024 scoping review of 69 studies found that 41 studies (59%) showed improvements in student academic achievement and 25 studies (36%) showed improvements in students' engagement. A 125 study synthesis from the systematic review in the education sectors in 2025 found that adaptive learning, machine learning, natural language processing, recommendation systems, intelligent tutoring, and generative AI are key technical methods to personalisation. There were 29 empirical studies chosen for a 2025 meta-analysis of AI and academic outcomes, and 99 effect sizes from 32 RCTs used in a separate meta-analysis for the effects of AI for personalised STEM education. This suggests that, alongside automation, academic benefits for AI depend on adaptive feedback in instruction, diagnostic assessment, appropriate sequencing in instruction, control by the learner, involvement of the teacher, and pedagogical alignment. The paper presents a conceptual model that connects the concepts of AI-assisted Learner Profiling, Adaptive Content, Feedback, Engagement, Self-regulation, and Academic Achievement. The analysis also highlights data privacy, unequal access, algorithmic limitations, educator readiness, and over-reliance on automated outputs as conditions that need to be protected within institutional contexts. The results encourage a human-centred approach in which AI is used as an adaptive teaching layer, not to replace teachers.
The topics include, but are not limited to, the following: artificial intelligence, personalised education, adaptive learning, academic performance, intelligent tutoring systems, learning analytics, student engagement, educational technology.