Generative AI in Higher Education: A Systematic Review of Opportunities, Challenges, and Pedagogical Innovations (2022–2025)

Synthesis: This PRISMA-guided systematic review synthesizes 125 peer-reviewed studies (2022–2025) on generative AI in higher education, documenting exponential adoption (92% student usage by 2025), four primary application domains, and persistent challenges around academic integrity, bias, hallucination, faculty readiness, and digital equity.

Key Findings

1. 125 studies, PRISMA 2020. The review synthesizes empirical evidence across 125 peer-reviewed studies following PRISMA 2020 guidelines, covering Scopus, Web of Science, ERIC, IEEE, and Google Scholar.

2. Exponential GenAI adoption. Student GenAI usage reached 92% by 2025, reflecting exponential growth since ChatGPT's release in November 2022.

3. Four primary application domains. Applications cluster into automated feedback and assessment, personalised learning support, critical skill development, and research assistance.

4. Balanced benefits and challenges. Benefits include enhanced learning efficiency, accessibility, and personalisation; challenges include academic integrity, algorithmic bias, hallucination rates, faculty readiness gaps, and digital equity concerns.

Implications

This broad systematic review provides a comprehensive landscape of Generative AI in Higher Ed, useful as a synthesis anchor for the wiki's coverage of GenAI applications. The 92% student-usage figure and the four-domain taxonomy (Automated Assessment, Personalized Learning, critical skills, research assistance) offer a useful organizing structure.

The persistent challenges — Academic Integrity, Bias Mitigation/algorithmic bias, Hallucination Risk, faculty readiness, and Equity/Equity — echo themes across the wiki and connect to AI Literacy and Faculty Development needs. The review's identification of pedagogical innovations (new assessment designs, AI literacy curricula, human-AI collaborative models) aligns with Human AI Collaboration and Human In The Loop AI threads.

For policy and practice, the review underscores that the faculty readiness gap and equity concerns remain the binding constraints on responsible GenAI integration, connecting to Educational Policy AI and AI Governance Education.