The Governance Gap Index

This book is the framework volume for an index. The Governance Gap Index is a structured way to compare what 195 nations have written down about artificial intelligence - laws, strategies, policies, ministerial declarations, agency guidance - against what those nations have actually built in the way of enforceable capacity to govern AI systems used inside their borders. It is not a ranking of which countries are "best" at AI. It is not a forecast of which countries will win an AI race. It is not an endorsement of any single regulatory model. It is a measurement instrument, and like any measurement instrument it is only as useful as the honesty with which we describe its limits. The present volume does not publish a complete 195-country scorebook. It defines the dimensions, source classes, coding rules, confidence labels, and responsible-use cautions needed before such a scorebook would be defensible. Chapter 4 includes an illustrative ten-jurisdiction matrix to show how the coding grammar works on real public-source records, but the full country-by-country release is deliberately deferred until every jurisdiction can carry a visible evidence trail. That scope choice is conservative rather than evasive: a governance index is only useful if readers can audit the cells behind the numbers. The premise of the index is uncomfortable: in most jurisdictions, what a government has said about AI and what a government can do about AI are not the same thing, and the distance between the two is itself a policy-relevant variable. A national AI strategy posted to a ministry website is one kind of artifact. A statute with binding effect, an empowered regulator with a budget, a procurement rule that actually constrains how the state buys AI systems, a pathway for citizens to obtain redress - those are different kinds of artifacts. They live in different parts of the governance stack, they require different resources to produce, and they fail in different ways. Conflating them produces the false impression that governance is further along than it is. The Organisation for Economic Co-operation and Development's policy observatory documents hundreds of national AI policy initiatives across its member and partner economies, but the observatory itself is careful to catalogue initiatives rather than to certify their effect (Organisation for Economic Co-operation and Development 2025). UNESCO's Recommendation on the Ethics of Artificial Intelligence, adopted by member states in 2021, lays out normative principles and explicitly contemplates a readiness assessment methodology because the framers anticipated that adopting principles is not the same as implementing them (UNESCO 2021). The European Union's Artificial Intelligence Act, Regulation (EU) 2024/1689, is one of the few instances in which a major jurisdiction has put a binding, risk-tiered statute into force, and even there the question of enforcement capacity - staff, technical expertise, market-surveillance authority - remains an active subject of public debate (European Union 2024). The Governance Gap Index treats these differences as data, not as embarrassments to be smoothed over. There are already serious comparative measures of national AI activity. Oxford Insights publishes the Government AI Readiness Index, which scores governments on dimensions including vision, governance, digital capacity, infrastructure, and skills (Oxford Insights 2024). The International Monetary Fund maintains an AI Preparedness Index that emphasizes the macroeconomic and institutional conditions under which AI adoption proceeds (International Monetary Fund 2024). Stanford's Institute for Human-Centered Artificial Inte