AI-enhanced creative pedagogy: A systematic review of AI interventions in fostering students’ creativity across digital humanities contexts

Abstract

Traditional pedagogical approaches systematically constrain creative thinking development through teacher-directed environments, standardized assessments, and implicit rather than explicit attention to creativity outcomes. This systematic review synthesizes empirical evidence on the effectiveness of artificial intelligence interventions in fostering students' creativity across digital humanities contexts. Following PRISMA 2020 guidelines, comprehensive searches of Web of Science, Scopus, ERIC, PsycINFO and IEEE Xplore identified 19 studies involving 1,695 participants across elementary through higher education contexts. Included studies employed experimental and quasi-experimental designs that examined AI tools—including large language models, visual generative systems, and specialized platforms—integrated into constructivist pedagogical frameworks. Meta-analytic findings demonstrate that AI interventions consistently produce large, statistically significant effects on creative thinking, with Cohen's d values exceeding 1.0 in multiple contexts. Fluency exhibited the most significant advancements across all educational tiers, whereas enhancements in originality and elaboration were dependent on structured scaffolding protocols. However, concerns regarding cognitive dependency, authenticity erosion, and academic integrity emerged consistently across studies. Assessments of publication bias showed that there were manageable threats to validity, and the fail-safe N analysis showed strong results. Effectiveness depends fundamentally on pedagogical orchestration positioning AI as a reflective partner rather than an autonomous replacement, with teacher competence and ethical frameworks proving decisive.