Regional governance: Tiered sovereignty

A group led by Data Science Nigeria has put out a new artificial intelligence policy paper for 19 countries across Africa, the Middle East, and Türkiye, and its central idea is that sovereignty does not have to be all-or-nothing. The paper says countries can keep tighter control over their most sensitive data and systems while still joining regional and global artificial intelligence networks where sharing is useful. (globalcenter.ai) That is a direct answer to a problem many governments now face: the best artificial intelligence systems improve with large pools of data, cloud access, and cross-border collaboration, but privacy law, national security, and political mistrust push in the opposite direction. The report frames data protection, interoperability, accountability, auditing, and cloud concentration as the pressure points that keep showing up across the region. (globalcenter.ai) The region in this paper is called Africa, the Middle East, and Türkiye, shortened to AMET, and the authors say its governments are not starting from one common legal model. They describe four broad starting points for national policy: security-led systems, innovation-led systems, rights-based systems, and development-focused systems. (globalcenter.ai) That matters because a country building artificial intelligence for border control does not regulate like a country building it for startup growth, and neither of those looks like a country focused on public services. The paper says these different incentives create “recognisable patterns of regulation” across the region rather than one uniform rulebook. (globalcenter.ai) The new phrase in the report is “tiered sovereignty,” which works like customs lanes at an airport. Some data and systems stay in the red channel under strict local control, while lower-risk uses can move through faster shared channels with standards, audits, and agreed safeguards. (globalcenter.ai) The paper ties that idea to concrete infrastructure choices, not just legal theory. It points to regional cloud centers, culturally calibrated testing layers, and shared standards as ways to let countries cooperate without giving up control over everything at once. (globalcenter.ai) This is also a sign of how the artificial intelligence debate is shifting outside the United States and Europe. A 2026 Brookings report says governments are treating artificial intelligence sovereignty as a question of infrastructure, data, and strategic autonomy, while a World Economic Forum paper says inclusive global governance now depends on interoperability rather than one universal model. (brookings.edu) (weforum.org) The report’s authors are not arguing for isolation. They explicitly recommend “flexible alignment,” cooperation through standards, and shared mechanisms that reduce fragmentation while accommodating political diversity across the region. (globalcenter.ai) Meta’s name shows up around this ecosystem in two ways. Meta publicly promotes responsible use, privacy, security, transparency, and governance for large language models, and the Global Center on AI Governance lists Meta among its funded activities in its 2024–2025 annual report, which supports the claim that this framework has backing from the company even though the report itself is published by the regional research network rather than by Meta. (ai.meta.com) (globalcenter.ai) What this paper is really trying to do is give governments a middle setting between two bad options: total openness that weakens local control, and total lockdown that cuts them off from the scale modern artificial intelligence needs. In a region where legal capacity, cloud access, and geopolitical leverage vary sharply from country to country, a tiered model is a way to say not every dataset, model, or public system has to cross the border under the same rule. (globalcenter.ai)