Bridging the AI Divide: A Reflexive Governance Framework for Small AI Adoption in Civil-Society Organisations

Two years after the adoption of the Global Digital Compact, the gulf between the AI governance capacity of the Global North and the Global South has not narrowed; in several measurable dimensions, it has widened. The Global AI Readiness Index records a composite governance score of 55 for high-income countries against 27 for developing economies, a two-to-one disparity that mirrors and amplifies the underlying connectivity, skills, and investment divides. The Independent International Scientific Panel on AI delivered its preliminary assessment in July 2026, and the first Global Dialogue on AI Governance concluded in Geneva with over fifteen hundred written submissions, yet the institutional mechanisms established by the Compact remain populated predominantly by Northern expertise and shaped by Northern epistemic frameworks. This scholarly article, published by the CyberSec Journal in collaboration with AUN Digital, argues that bridging the AI divide requires not only new indices and capacity-building programmes but a fundamental shift in how civil-society organisations participate in and model AI governance. Drawing on the concept of reflexive governance, the report introduces a framework whereby advocacy organisations that adopt AI tools in their own operations can generate credible, citable evidence about responsible AI use in development contexts, thereby contributing to the very governance debate they seek to influence. The report surveys the post-Compact institutional landscape, assesses the coverage gaps in existing measurement frameworks, and proposes a Reflexive Governance Framework for civil-society AI adoption, calibrated to the forward runway leading to the 2027 Global Dialogue in New York and the high-level review of the Compact at the eighty-second session of the General Assembly.