Artificial intelligence in higher education: a systematic review of its impact on student engagement and the mediating role of teaching methods

Synthesis: Long, Wang, Md Rashid, and Lu (2026) systematically review 73 peer-reviewed studies (2015–early 2025, Scopus and Web of Science) on AI in higher education and its impact on student engagement, with a focus on the mediating role of teaching methods. They find that AI tools — chatbots, adaptive systems, predictive analytics — enhance engagement most effectively when embedded within interactive pedagogies (flipped classrooms, project-based learning, scaffolded feedback loops). They introduce the PMAISE model (Pedagogical Mediation of AI for Student Engagement), mapping the alignment between AI technologies, pedagogical strategies, and the affective, behavioral, and cognitive dimensions of engagement, and examine ethics, data privacy, and structural barriers to equitable adoption.

Key Findings

1. 73 studies (2015–early 2025), PRISMA-guided. A systematic review from Scopus and Web of Science, coded by AI type, engagement outcomes, and instructional strategies.
2. Pedagogy mediates AI's effect on engagement. AI enhances engagement most effectively within interactive pedagogies — flipped classrooms, project-based learning, and scaffolded feedback loops — rather than through tool deployment alone.
3. PMAISE model introduced. The review proposes a conceptual framework (Pedagogical Mediation of AI for Student Engagement) aligning AI technologies, pedagogical strategies, and the affective, behavioral, and cognitive dimensions of engagement.
4. Teaching methods amplify or inhibit AI effects. Concrete examples demonstrate that the same AI tool can boost or fail to boost engagement depending on the surrounding instructional design.
5. Equity and ethics concerns. Structural barriers to equitable AI adoption (e.g., socioeconomically disadvantaged students more vulnerable to unstable internet or limited device access), alongside data privacy and ethical concerns, are examined.

What this means for practice

• Instructors. Embed AI tools inside interactive pedagogies — flipped classrooms, project-based learning, and scaffolded Feedback loops — rather than adding them as standalone resources, because the teaching method mediates whether engagement improves.
• Designers. Use the PMAISE model to check alignment between the AI tool, the pedagogical strategy, and the affective, behavioral, and cognitive dimensions of engagement before committing to a tool.
• Instructors. Plan the pedagogy before the tool: the review shows the same AI tool can raise or fail to raise engagement depending on the instructional design around it.
• Administrators. Address structural barriers alongside the tool rollout — unstable internet and limited device access fall hardest on socioeconomically disadvantaged students — and pair deployment with data privacy safeguards.

Limitations

• As a systematic review of 73 articles retrieved from Scopus and Web of Science for 2015 to early 2025, the findings are bounded by those two databases, the PRISMA screening criteria, and the quality of the included studies.
• The PMAISE model is a conceptual contribution proposed by the authors, not an empirically validated framework, so it should be treated as a hypothesis about pedagogical mediation rather than as tested structure.
• Causal claims about engagement are limited by the designs of the underlying studies, which include observational work alongside intervention studies.

Citation

Long, D. Y., Wang, S., Md Rashid, S., & Lu, X. T. (2026). Artificial intelligence in higher education: a systematic review of its impact on student engagement and the mediating role of teaching methods. Frontiers in Education, 10, 1648661. https://doi.org/10.3389/feduc.2025.1648661