Trend Regulators worldwide shift to risk-based and experimental rules for emerging technologies

Summary

International practice is moving toward risk-based regulation and regulatory experimentation to cope with the uncertainty of emerging technologies such as AI, synthetic biology, autonomous driving and quantum. The EU AI Act, in force since August 2024, introduced the world's first legally binding risk-based AI regime, scaling obligations to each system's risk rather than applying uniform rules. The OECD's 2025 regulatory outlook promotes agile regulatory governance built on strategic intelligence, similar to horizon scanning and foresight, and continuous learning. Its July 2025 sandbox toolkit defines sandboxes as tools for understanding the opportunities and risks of specific innovations and building suitable rules. The EU now requires all 27 member states to set up AI regulatory sandboxes by 2026, and Spain selected 12 pilot projects in 2025 as the first case.