The World's Strictest Law on Human-Like AI
China's new law on anthropomorphic AI comes into effect tomorrow. Here is what it says | Edition #306
TL;DR
- China's new law on anthropomorphic AI, effective July 15, 2026, is the world's strictest and most comprehensive.
- The law applies to AI services that simulate human personality characteristics for continuous emotional interaction.
- Providers must ensure national security, protect public interests, and safeguard citizens' rights.
- Prohibited activities include generating content that endangers national security, incites harm, or infringes on rights.
- Service providers are responsible for security, including algorithm and ethics review, and risk contingency plans.
- Measures are in place to protect minors, including consent requirements and a 'minor mode'.
- Providers must identify AI-generated content and notify users, and implement exit strategies for services.
- Security assessments are required for new services, significant changes, or services with over 1 million registered users.
- Penalties for violations range from warnings to service cessation and substantial fines.