In CAC's rules, AI safety means orderly information services: security assessment and filing before public deployment, traceability and labeling of synthetic content, and provider responsibility for content risk.
Cyberspace Administration of China
国家互联网信息办公室
Every public-facing generative AI service in China clears rules CAC issued or co-issued; its filing, security-assessment, and labeling architecture is the enforcement layer the rest of the stack builds on.
Importance
Assessed per aspect, band 0 to 3. There is no composite score.
CAC issued or co-issued the four binding rules that govern algorithmic recommendation, deep synthesis, generative AI services, and AI content labeling.
The 2024 national AI standardization guide lists the Central Cyberspace Affairs Commission Office among its joint issuers, and the mandatory labeling standard GB 45438-2025 records CAC as its administering authority.
What safety means here
Leadership
No leadership roles are recorded with a verifiable source.
History
- 2021-12-31
Algorithmic recommendation provisions published as order No. 9, jointly with MIIT, MPS, and SAMR.
- 2022-12-11
Deep synthesis provisions posted, jointly issued with MIIT and MPS.
- 2023-07-13
Interim measures for generative AI services published.
- 2024-07-03
National AI standardization guide (2024 edition) posted.
- 2025-03-14
AI-generated content labeling measures published, effective 2025-09-01.
Connections
Recorded edges are co-issuance links with MIIT; each of the four binding instruments is a joint document rather than a solo CAC order. analyst judgment
Co Issues
- Ministry of Industry and Information Technology CAC and MIIT are both named co-issuing ministries for the Interim Measures for the Management of Generative AI Services.
- Ministry of Industry and Information Technology CAC and MIIT are both named issuers of the Deep Synthesis Provisions, together with MPS.
- Ministry of Industry and Information Technology CAC and MIIT are both named issuers of the AI-generated content labeling measures, together with MPS and NRTA.
- Ministry of Industry and Information Technology CAC's parent office and MIIT jointly issued the 2024 national AI industry comprehensive standardization guide.
Selected outputs
- Interim Measures for the Management of Generative AI Services2023
- Measures for Labeling AI-Generated and AI-Synthesized Content2025
- Provisions on the Administration of Algorithmic Recommendation in Internet Information Services2021
- Provisions on the Administration of Internet Information Services Deep Synthesis2022
Engagement fit
- Plausible for
- Formal, regulator-to-regulator exchange on service rules, filing procedure, and labeling implementation, if framed through official channels.
- Constrained by
- It is a rule-issuing and enforcement body; open technical exchange on model evaluation sits more naturally with the standards and institute layer below it.
Open questions
- Whether to model the Central Cyberspace Affairs Commission Office and CAC as one node or as linked but distinct nodes.
- Current leadership could not be recorded: the reference naming CAC's director did not resolve to a verifiable source.
Source appendix (7 records)
- cac-algorithm-recommendation-provisions-2022
Provisions on the Administration of Algorithmic Recommendation in Internet Information Services · Cyberspace Administration of China · 2021-12-31open
- cac-deep-synthesis-provisions-2022
Provisions on the Administration of Internet Information Services Deep Synthesis · Cyberspace Administration of China · 2022-12-11open
- cac-generative-ai-measures-2023
Interim Measures for the Management of Generative AI Services · Cyberspace Administration of China · 2023-07-13open
- cac-ai-content-labeling-measures-2025
Measures for Labeling AI-Generated and AI-Synthesized Content · Cyberspace Administration of China · 2025-03-14open
- cac-national-ai-standardization-guide-2024
Notice on Issuing the National Artificial Intelligence Industry Comprehensive Standardization System Construction Guide 2024 Edition · Cyberspace Administration of China · 2024-07-03open
- gb-45438-2025-aigc-labeling-standard
GB 45438-2025 Cybersecurity Technology: Labeling Method for AI-Generated and Synthetic Content · State Administration for Market Regulation / Standardization Administration of China · 2025-02-28open
- report-china-standards-evaluation-2026-06-05
China Standards and Evaluation Research Brief · Asia AI Safety Atlas research brief
Last verified 2026-07-07