EU AI Act
Risk-based horizontal regulation with prohibited practices, GPAI obligations, transparency rules and high-risk requirements. Important application dates are phased, with enforcement powers applying from 2 August 2026.
Global AI Governance · Updated September 2026
Global
AI Acts.
There is no single global AI law. Europe has a horizontal AI Act. The United States uses a distributed federal, state and sectoral approach. China relies on layered binding measures. Other jurisdictions are building their own models.
Distributed AI governance
No single federal horizontal AI Act currently mirrors the EU model. Federal executive policy, agencies, sectoral law and state legislation form a distributed framework.
Layered binding measures
China regulates recommendation algorithms, deep synthesis, generative AI services and AI-generated synthetic content. AI content identification measures took effect on 1 September 2025.
Sector-led framework
The UK primarily uses existing regulators and sectoral rules. The Artificial Superintelligence Bill is a parliamentary proposal, not current law.
Promotion + governance
Japan’s approach emphasizes development, utilization, national strategy and risk governance with a comparatively innovation-oriented model.
AI Basic Act
The current amended Framework Act on AI Development and Trust is effective 21 July 2026 and includes provisions for high-impact AI, safety, trust and oversight.
This page is an informational overview and not legal advice. AI regulation changes quickly; always consult the latest official text and competent legal counsel for compliance decisions.
Verification layer
Official
sources.
European Commission · AI ActOpen source ↗
European Commission · AI Act enforcementOpen source ↗
United States · White House AIOpen source ↗
China · Cyberspace AdministrationOpen source ↗
United Kingdom · Parliament BillsOpen source ↗
Japan · Cabinet Office AIOpen source ↗
South Korea · Official law databaseOpen source ↗