On July 17, 2026, the research report The Differentiated Global South: Capability Gaps and Strategic Pathways in AI Governance (hereinafter referred to as "the Report"), produced by the Center for Global AI Innovative Governance, was officially released at the forum themed "Global AI Governance and Sustainable Development: Challenges and Responses" during the 2026 World Artificial Intelligence Conference (WAIC 2026).
As one of the deliverables of the Fudan Report Series, the Report is co-launched by the Center for Global AI Innovative Governance and the Institute of International Studies at Fudan University. It is co-authored by Cai Cuihong, Professor at the Institute of International Studies of Fudan University and Deputy Director of the Center for Global AI Innovative Governance, and Guan Hang, doctoral candidate at the School of International Relations and Public Affairs of Fudan University.
The Report points out that when global AI governance is framed as competition between major tech powers such as China and the United States, a fundamental question is often overshadowed: what role do Global South countries, which account for the majority of the world’s population, actually play in this sweeping process of institutional building?The Report conceptualizes the Global South as a political-economic space that is being re-stratified by gaps in AI capabilities. Four asymmetries inherent in the structure of global AI governance collectively shape the internal differentiation of the Global South. Among these, the asymmetry in infrastructure distribution among Global South countries sets the threshold for AI capacity building, while the asymmetry in the distribution of capital and models shapes power dynamics and dependency relations between the Global South and external actors. Beyond the Global South, the asymmetry in the patterns of rule spillovers from leading tech powers and international organizations amplifies domestic pressure for institutional adaptation. As for the asymmetry in the influence of multilateral mechanisms, while it offers options and operational space for the Global South to take action, it also objectively constrains their agenda-setting capacity and their ability to translate policy agendas into concrete actions.
The Report constructs a two-dimensional “capacity–normative orientation” matrix. It measures what a country “can do” across dimensions including technological infrastructure, innovation and R&D, capital and industrial ecosystem, talent and skills, and institutional and regulatory capacity, and forms a comprehensive judgment on where a country “aims to head” by synthesizing its national artificial intelligence strategies, draft legislation, budgetary allocationsand orientation of international cooperation. On this basis, countries of the Global South are categorized into four governance archetypes: the development-priority archetype represented by most member states of the African Union (AU), the rule-balancing archetype represented by the Association of Southeast Asian Nations (ASEAN), the sovereignty-shaping archetype represented by India, Brazil, the United Arab Emirates (UAE) and Saudi Arabia, and the externally embedded archetype represented by Pacific island countries and Caribbean states.
The Report observes that under multiple structural constraints, the Global South can hardly evolve rapidly into an institutionalized "third pole" in global AI governance. A more realistic near-term trend is the formation of loose issue-based coordination around specific topics including computing power affordability, open-source AI models, local language data resources, and AI for public services.At the policy level, the Report proposes reintegrating the Global South into the institutional origins of AI governance. The international community should establish an AI Capacity and Development Summit mechanism, jointly set up a multilateral financing window for capacity building, promote mutual recognition of regional regulatory sandboxes and testing tools, embed "non-lock-in" contract clauses in external cooperation, and roll out phased governance pathways in countries with weak AI capacity.
The Report stresses that inclusive governance is ultimately anchored not only in representation in decision-making but also in capacity provision. It is no difficult task to include developing countries on the list of signatories to relevant governance instruments. What truly embodies structural inclusion, however, is equipping them with genuine capabilities to evaluate AI models, exercise regulatory oversight over AI procurement, develop datasets, and utilize computing power. Only when capacity provision becomes an integral part of the global AI governance institutional framework can the Global South move from the margins of rule-making to the institutional origins of the system.
About the author
Cai Cuihong, Deputy Director of the Center for Global AI Innovative Governance and Professor at the Institute of International Studies of Fudan University;
Guan Hang, doctoral candidate at the School of International Relations and Public Affairs of Fudan University.
Original URL: 
The Differentiated Global South: Capability Gaps and Strategic Pathways in AI Governance.pdf
