Synergizing Artificial Intelligence and Multiple Intelligences in Project-Based Learning: A Meta-Analysis of Academic Achievement Outcomes

Abstract

Project‑based learning (PBL) promotes deeper learning but often fails to accommodate diverse cognitive profiles. Artificial intelligence (AI) offers adaptive scaffolding, while Multiple Intelligences (MI) theory provides a differentiation framework. However, no quantitative synthesis has examined their combined effect on academic achievement. This meta‑analysis synthesizes evidence on the synergy of AI and MI within PBL and estimates the overall effect on academic achievement, together with key moderators. Methods: Following PRISMA 2020 guidelines, we systematically searched Scopus, Web of Science, ERIC, PsycINFO, and ProQuest Dissertations (2015–2025). Inclusion criteria were: (a) empirical studies with control/comparison groups; (b) interventions combining AI tools and MI‑based differentiation in PBL; (c) reported academic achievement data; (d) K‑16 learners. Random‑effects meta‑analysis, moderator analyses, and publication bias tests were performed. Forty‑two studies (N = 8,943) were included. The overall effect was moderate and positive (Hedges’ g = 0.48, 95% CI [0.39, 0.57], p < .001). Significant moderators were AI role (scaffolding > content generation > assessment), MI implementation method (student choice > teacher‑assigned/fixed), and education level (secondary > primary > tertiary). Subject domain did not moderate the effect. Publication bias was minimal (Egger’s p = 0.12), and sensitivity analyses confirmed robustness. AI and MI synergize effectively in PBL, yielding meaningful academic gains that exceed the isolated effects of either component. Educators should embed AI as a scaffolding tool (not an automaton) and allow students to choose MI‑aligned project roles. Policymakers should invest in AI tools with MI‑differentiation capabilities for PBL curricula.