National Standard of the People’s Republic of China: Cybersecurity Technology – Basic Safety Requirements for Generative Artificial Intelligence Services

• Authority: Chinese central government
• Jurisdiction: China
• Collections: Chinese law and policy
• Most recent activity: Proposed
• Most recent activity date: 2025-04-25

Risk subdomain coverage

Good coverage — Comprehensive governance measures or mitigations with clear procedures relating to the risk subdomain are explicitly described.

• Exposure to toxic content
• Compromise of privacy
• AI system security vulnerabilities
• Lack of capability or robustness
• Lack of transparency or interpretability

Minimal coverage — Brief mention but no specific mitigations, recommendations or governance measures relating to the risk subdomain are described.

• False or misleading information
• Disinformation, surveillance, and influence at scale
• Cyberattacks, weapon development or use
• Overreliance and unsafe use
• Economic and cultural devaluation of human effort
• AI pursuing its own goals
• AI possessing dangerous capabilities

All other risk subdomains had no coverage in this document.

Risk Coverage Summary

The document has good coverage of approximately 8-10 subdomains, with strong focus on toxic content (1.2), privacy (2.1), security vulnerabilities (2.2), false information (3.1), lack of robustness (7.3), and lack of transparency (7.4). Coverage is concentrated in security, content safety, and operational reliability domains, with minimal attention to socioeconomic impacts or advanced AI risks.

Sector coverage

Good coverage — Comprehensive governance measures or mitigations with clear procedures relating to the sector are explicitly described.

• Information

Minimal coverage — Brief mention but no specific mitigations, recommendations or governance measures relating to the sector are described.

• Trade, Transportation & Utilities
• Finance & Insurance
• Educational Services
• Health Care & Social Assistance
• Public Administration (excl. National Security)
• National Security

All other sectors had no coverage in this document.

Sectors Summary

The document governs Generative AI services across all sectors but provides explicit additional requirements for high-risk applications including critical information infrastructure, healthcare (medical information services, psychological counseling), financial information services, public administration (social governance, public security), and services for minors. The standard is cross-sectoral with heightened obligations for sensitive domains.

Legislative Status

Hard Law

Legislative Status Summary

This is a National Standard of the People's Republic of China (GB/T series), issued by the Chinese central government. The document uses predominantly mandatory language ('shall', 'must') throughout and establishes binding requirements for GenAI service providers with explicit oversight by government departments and enforcement through third-party evaluation institutions.

Actors

Proposer

• Entity Types: AI_Governance_Actor
• Entities: Chinese central government; National standardization authority (implied by GB/T designation)
• Summary: The document is identified as a 'National Standard of the People's Republic of China' with a GB/T designation, indicating it was developed and issued by China's national standardization system under central government authority.

Target

• Entity Types: AI_Developer, AI_Deployer
• Entities: Service providers conducting activities related to GenAI services
• Summary: The document explicitly states it 'applies to service providers conducting activities related to GenAI services.' Service providers are comprehensively defined as organizations or individuals providing GenAI services through various interfaces, covering both development and deployment functions.

Enforcer

• Entity Types: AI_Governance_Actor
• Entities: Main oversight departments (主管部门); Third-party evaluation institutions
• Summary: The document explicitly identifies 'main oversight departments' as the primary enforcement bodies and mentions third-party evaluation institutions as part of the enforcement and compliance verification system. The document is designed to support filing, registration management, testing, and evaluation activities by these entities.

Monitor

• Entity Types: AI_Governance_Actor
• Entities: Third-party evaluation institutions; Main oversight departments (主管部门); Monitoring personnel within service provider organizations
• Summary: Third-party evaluation institutions are explicitly mentioned as playing a monitoring/assessment role. The document also requires service providers to establish internal monitoring personnel and mechanisms for ongoing compliance verification and safety monitoring.

AI Lifecycle Stages

Lifecycle stages covered by the documents, as defined in the OECD Framework for the Classification of AI Systems

• Plan and Design
• Collect and Process Data
• Build and Use Model
• Verify and Validate
• Deploy
• Operate and Monitor

AI Lifecycle Summary

The document comprehensively covers the full AI lifecycle with particularly strong emphasis on data collection and processing, model building and training, verification and validation through testing, deployment requirements, and extensive operational monitoring. The document focuses specifically on Generative AI services.

Technical Scope

Terms relating to technical scope that are mentioned or implicitly referenced in the document

• AI Models
• Foundation Models
• AI Systems
• Generative AI
• Open Weight / Open Source

Technical Scope Summary

The document explicitly and primarily focuses on Generative Artificial Intelligence (GenAI) throughout, with clear definitions. It addresses both AI models and AI systems, though models receive more attention. There is no mention of frontier AI, general purpose AI, task-specific AI, foundation models, predictive AI, open-weight models, or specific compute thresholds.