AI, Structural Power and Governance Fragmentation: A Systematic Review and the Geopolitical AI Power Framework

Debates on artificial intelligence and international politics are dominated by the rivalry between the United States and China, which leaves out how other countries become dependent on infrastructures, models and rules they do not control. This article asks how AI reconfigures structural power in the global political economy and why it fragments its governance. It reports a systematic review following PRISMA 2020: 1,763 records from Scopus and the Web of Science Core Collection (2018–2026) were deduplicated to 1,199 and screened by two independent automated screeners (rule-based and zero-shot natural language inference; κ = 0.457), with adjudication of conflicts. Of these, 330 studies were mapped through keyword co-occurrence, and a core of 49 informed a qualitative synthesis. The literature forms five clusters, led by sovereignty and the Global South and by great-power regulatory competition, while work on data extraction remains dispersed. Building on Strange, Gilpin and Zuboff, the article proposes the Geopolitical AI Power Framework (GAPF), in which structural power asymmetry, techno-nationalist fragmentation and surveillance capitalism externalities reinforce one another and produce structural lock-in. Lock-in is a tendency, not a destiny: semi-peripheral countries can still shape their position through capabilities, negotiation and regional coordination.