⚡ Quick Summary
This book offers one of the most rigorous and conceptually mature treatments of global AI governance to date. Matthijs M. Maas argues that AI governance cannot be understood or designed through isolated legal instruments or single institutions. Instead, it must be approached as a complex, evolving governance architecture shaped by overlapping regimes, competing actors, and rapid sociotechnical change. The work systematically examines whether effective global AI regulation is feasible and concludes that it is—not because governance is simple, but because enforceable policy levers already exist if coordination can be achieved. The book moves the debate beyond abstract calls for treaties or agencies and toward a structured way of thinking about how governance architectures actually form, fragment, and adapt over time. It is both a diagnostic map of the current global AI governance landscape and a strategic toolkit for shaping its future.
🧩 What’s Covered
The book is structured into three interlocking parts that build from foundations to practical choice.
First, it establishes the foundations of global AI governance. This includes defining AI governance as a field, situating it within international law and global governance theory, and explaining why AI creates collective action problems that cannot be solved at the national level alone. Maas carefully dismantles the assumption that global AI governance is infeasible, showing instead that enforcement barriers are often political rather than technical. He reviews existing governance levers, from export controls and standards to compute governance and institutional coordination.
Second, the book introduces three conceptual lenses for understanding change. The sociotechnical change lens reframes AI not as a static technology but as a driver of shifting behaviors, affordances, and societal impacts, which in turn create distinct regulatory rationales. The governance disruption lens explores how AI can undermine, displace, or destabilize existing legal and institutional arrangements, including rulemaking, enforcement, and legitimacy. The regime complexity lens then ties these threads together, explaining how AI governance emerges as a regime complex: a dense web of partially overlapping institutions, norms, and actors whose interactions can be both productive and dysfunctional.
Third, the book provides a practical analytical framework for shaping global AI governance. Maas proposes a step-by-step method for analyzing and designing AI regime complexes, focusing on origins, topology, evolution, consequences, and management strategies. This framework helps readers evaluate whether fragmentation, centralization, or hybrid approaches are likely to improve coherence, resilience, and effectiveness. Throughout, the book draws on historical analogies, contemporary policy initiatives, and emerging proposals such as international AI agencies, compute caps, and certification regimes.
💡 Why it matters?
This book matters because it replaces wishful thinking with structural realism. It explains why global AI governance currently feels chaotic, why fragmentation is not necessarily failure, and how governance architectures can be deliberately shaped rather than passively endured. For anyone working on the EU AI Act, international AI coordination, safety governance, or institutional design, it provides a shared conceptual language and a disciplined way to reason about trade-offs. It also bridges law, international relations, and technology policy in a way that few AI governance works manage to do convincingly.
❓ What’s Missing
The book is intentionally analytical rather than prescriptive. Readers looking for concrete institutional blueprints, model treaty text, or operational governance playbooks may find the conclusions demanding rather than directive. There is also limited engagement with day-to-day regulatory implementation challenges faced by national authorities or companies. The strength of the framework lies in diagnosis and strategic orientation, not in offering ready-made policy shortcuts.
👥 Best For
AI governance researchers, international lawyers, policy designers, and advanced practitioners working on cross-border AI regulation. It is especially valuable for those involved in international coordination, institutional design, or long-term AI safety and governance strategy.
📄 Source Details
Architectures of Global AI Governance: From Technological Change to Human Choice
Matthijs M. Maas
Oxford University Press, 2025
📝 Thanks to
Matthijs M. Maas for delivering a rare combination of theoretical depth, interdisciplinary clarity, and practical relevance in the global AI governance debate.
