10 min read
Lesson 3 of 5
Global AI Governance Approaches
From the course AI Regulation & Policy Landscape
Beyond the EU and US, countries worldwide are developing distinct approaches to AI governance reflecting their values, priorities, and development goals.
China's AI Regulations
China has moved quickly with targeted AI regulations:
- Algorithm Recommendation Regulation (2022) — transparency requirements for recommendation algorithms
- Deep Synthesis Regulation (2023) — deepfake labeling and registration requirements
- Generative AI Measures (2023) — content moderation, training data requirements, and registration for public-facing generative AI
- AI Safety Governance Framework — risk classification similar to the EU approach
United Kingdom
The UK pursues a pro-innovation, sector-led approach:
- No horizontal AI legislation — existing regulators apply AI-specific guidance
- AI Safety Institute — government institute for frontier AI safety evaluation
- Bletchley Declaration — hosted first AI Safety Summit in November 2023
- Regulatory Sandbox — allow AI innovation within supervised environments
Other Notable Frameworks
- Canada — Artificial Intelligence and Data Act (AIDA) as part of Bill C-27
- Brazil — comprehensive AI bill modeled partly on the EU AI Act
- Japan — voluntary guidelines emphasizing innovation and social benefit
- India — sector-specific approach with focus on responsible AI principles
- Singapore — Model AI Governance Framework with voluntary adoption
- South Korea — AI Basic Act establishing rights and obligations
International Coordination
Efforts to align global AI governance:
- G7 Hiroshima AI Process — guiding principles and code of conduct for AI
- OECD AI Principles — adopted by 46 countries, influencing national policies
- UNESCO Recommendation on AI Ethics — first global normative instrument
- UN AI Advisory Body — recommendations for global AI governance
- ISO/IEC 42001 — international standard for AI management systems
- Council of Europe AI Convention — first binding international AI treaty
Regulatory Divergence Risks
Different approaches create challenges:
- Companies must comply with multiple, sometimes conflicting frameworks
- Data localization requirements fragment global AI development
- Different risk classifications create compliance complexity
- Extraterritorial application of laws creates jurisdictional conflicts
Key Takeaways
- •China has targeted regulations for algorithms, deepfakes, and generative AI
- •The UK takes a pro-innovation approach through existing regulators
- •G7, OECD, UNESCO, and the UN are coordinating global AI governance
- •ISO/IEC 42001 provides an international AI management system standard
- •Regulatory divergence creates compliance challenges for global AI companies
0 of 5 lessons marked complete
Progress is kept in this browser only.