Effect of Artificial Intelligence tools on Academic Performance of students: A Meta Analysis

Author(s) : Zaheer Abbas 1* & Hafiz Tahir Jameel2

Abstract:

Artificial intelligence (AI) has gained rapid development in the field of education for improving classroom instruction, assessment and students’ learning outcomes. This meta-analysis specifically focused on the overall impact of AI on the academic performance of students. Using “Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA)” guidelines, the study synthesized 18 studies retrieved from three databases (“Directory of Open Access Journals - DOAJ, PubMed, and Google Scholar”) published between 2015 to June 2025 and used the information extracted from the studies for the analysis. The studies comprised a sample of 3436 participants from various educational levels fulfilled the inclusion criteria. For the analysis, the data from the studies was coded and entered into “Jaffery’s Amazing Statistical Program (JASP) version 0.18.3”. The results showed that there was a moderate to large (0.62) pooled effect size under the random-effects model. Moderating effects of four variables, type of institution, sample size, area of the subject and AI types were examined. Results depicted that the sample size moderated the effectiveness of AI in predicting the academic performance of students significantly. However, other moderating variables, that is, type of institution, area of the subject and AI type did not show any significant effects. Based on the findings, the meta-analysis presented practical implications for both policy and practice.

Keywords: Artificial Intelligence, Academic Performance, Moderators, Forest Plot, Funnel Plot.