Ethical and Inclusive Artificial Intelligence for Sustainable Knowledge Societies: A Global Governance Framework through a Systematic Literature Review

Purpose: To bridge the persistent gap between AI ethics principles and educational practice by proposing a Pedagogically-Centered AI Governance (PCAG) model that situates pedagogy as the generative core of ethical and inclusive artificial intelligence. Design/methodology/approach: A systematic literature review was conducted following PRISMA 2020 guidelines across seven databases (Scopus, Web of Science, IEEE Xplore, ScienceDirect, SpringerLink, Wiley, and Google Scholar) covering publications from 2014 to May 2024. Using the PICo framework for scoping and CIMO logic for synthesis, 127 peer-reviewed studies were analyzed to identify mechanisms linking ethical AI to inclusive educational outcomes. Findings: Four recurrent mechanisms were synthesized from the literature: (1) pedagogical capacity-building , (2) technical process control , (3) regulatory accountability , and (4) participatory legitimation . Integrating these mechanisms, the study proposes the Pedagogically-Centered AI Governance (PCAG) model and the Sustainable Educational Governance Framework (SEG-F) . Together, they conceptualize AI not merely as a technological tool but as an educational infrastructure that should be ethically and pedagogically governed to promote equity, inclusion, and epistemic justice. Practical implications: PCAG offers a multi-layered governance architecture and a six-item operational checklist for policymakers, educators, and designers to translate AI ethics into curriculum development, teacher training, and institutional decision-making. Originality/value: This review advances AI-in-education scholarship by combining PRISMA + PICo + CIMO into a mechanism-oriented synthesis and by centering pedagogy as the key lever for converting global AI ethics principles into contextually just and inclusive educational practices.