Global Frameworks, Local Realities: A Critical Review of Generative AI Governance in MENA Higher Education

Abstract

Generative AI has entered higher education faster than institutions can govern it, yet governance scholarship remains concentrated in the Global North. This structured, PRISMA-informed critical review of forty-nine sources examines how governance frameworks from international bodies (UNESCO), comparative national systems (the United States, Japan, China), and continental African and Maghreb/Algeria-specific strategies travel to under-resourced, multilingual, post-colonial contexts such as Algeria. The review's central finding is that governance frameworks built for high-resource, monolingual, well-capitalised systems are adopted by reference in contexts with markedly different linguistic, infrastructural, and institutional conditions. Evaluated against the criteria of linguistic inclusivity, infrastructural realism, and institutional capacity, the frameworks reviewed reveal three recurring cross-cutting gaps: a linguistic gap, an infrastructural-realism gap, and a capacity-and-voice gap. In response, the review proposes four governance adaptation principles for MENA higher education: linguistic parity, access-conditional design, enabling rather than restrictive framing, and locally deliberated governance. The contribution is a critical synthesis rather than new empirical data: it maps an underexamined literature, identifies specific gaps in how existing frameworks address multilingual and resource-constrained contexts, and offers these four principles as a starting agenda for national and institutional policy-makers in Algeria and comparable settings.