What’s Covered?
The report, titled The Anatomy of AI Rules, builds a comparative backbone for international AI governance. It analyzes 11 comprehensive AI rulebooks from seven jurisdictions: Argentina, Brazil, Canada, China, the EU, South Korea, and the US. These are dissected through a shared taxonomy based on the OECD AI Principles:
- Inclusive Growth
- Human Rights and Democratic Values
- Transparency and Explainability
- Robustness, Security, and Safety
- Accountability
The researchers mapped 74 regulatory requirements (e.g., data protection, non-discrimination, content watermarking) to these five principles and tracked their presence—or absence—across countries. This effort revealed a three-layered divergence:
- Priority divergence: Different principles matter more to different jurisdictions (e.g., China puts more weight on fairness; the EU and US emphasize accountability).
- Policy divergence: The same principle is implemented through different policy areas—data governance, consumer protection, content moderation, etc.
- Requirement divergence: Even when two countries use the same policy tool, like data protection, the legal specifics often don’t align.
The report includes granular analysis of how specific requirements are operationalized. For example, under Principle 1.2 (Human Rights), it looks at:
- Non-discrimination obligations: Whether systems must avoid biased outcomes, or users must be able to contest them.
- Data protection: Whether rules focus on user control, data security, or both.
- Human oversight and interaction rights: How people can control or understand AI systems they interact with.
It also includes detailed heatmaps comparing which countries have adopted which rules, allowing users to zoom in on gaps and overlaps with a few clicks via the CLaiRK interface.
💡 Why it matters?
This is a crucial tool for regulators, international bodies, and policy researchers trying to keep AI governance from spiraling into digital protectionism. It shows not only where we agree in principle—but where we fall apart in practice. In a world where AI crosses borders by default, interoperability will hinge on understanding these layers of divergence.
What’s Missing?
While the report offers razor-sharp comparative granularity, it doesn’t go deep into the why behind divergences. There’s little context about political economy, legal traditions, or industry lobbying pressures that shape these choices. There’s also limited practical guidance for harmonization—no roadmap, no best practices, no prioritization framework for closing gaps. It leaves readers well-informed but unequipped to act. Also, it doesn’t cover the implementation and enforcement phases, which are critical for moving from paper to practice. Some important jurisdictions (like Japan or India) are notably absent, limiting the global scope.
Best For:
Policy professionals, trade negotiators, and legal teams working in multinational contexts. It’s also a valuable resource for standard-setters and civil society actors pushing for alignment on global AI norms. Less useful for startups or product teams looking for implementation guidance.
Source Details:
Title: The Anatomy of AI Rules: A Systematic Comparative Analysis of AI Rules Across the Globe
Authors: Tommaso Giardini and Johannes Fritz, with contributions from Digital Policy Scholars at the Law and Economics Foundation, University of St. Gallen.
Publisher: Digital Policy Alert (DPA), in partnership with the Law and Economics Foundation, St. Gallen.
Context: DPA is a watchdog and tracking initiative aimed at surfacing regulatory changes across the digital economy. This report represents one of the most granular efforts yet to make AI rulebooks interoperable, backed by a custom taxonomy and tools like CLaiRK for live cross-jurisdictional navigation.
