China · Governance guidance

China: AI Safety Governance Framework

The framework supports graded risk identification and treatment across models, data, systems, networks, ethics and applications for AI development, deployment and use.

Current status
Current governance framework
Published or updated
15 September 2025
Last verified
19 August 2026

Prepared by Longan Bay Area AI Legal Research Center based on official public materials.

01

Scope of application & addressees

  • Organisations developing, training, supplying, deploying or using AI systems
  • Teams responsible for model, data, content and technology-ethics risks
  • Public bodies and businesses building an AI risk inventory and control programme

02

Core regulatory mandates & key requirements

01Risk levels reflect likelihood and impact and should be updated as technology and use change.

02The framework covers intrinsic, application, cognitive and ethical safety risks.

03Technical, organisational, coordinated-governance and international measures support risk treatment.

03

Enforcement & compliance timeline

Version 1.0 was released.

Version 2.0 was released.

Primary text

Statutory & regulatory text

PROJECT ENQUIRY

Need to evaluate this regulation against your system architecture?

Consult with our regulatory team →