China Measures for the Management of AI-Generated Content

Applies To

Overview

The Measures for the Management of AI-Generated Content were published by the Cyberspace Administration of China (CAC) and came into force in March 2025. They represent China's most specific regulatory instrument targeting the labelling and traceability of AI-generated text, image, audio, and video content, including synthetic media and deepfakes. The measures build upon and partly supersede earlier guidance issued under China's network information service regulations and are closely related to the Provisions on the Management of Algorithmic Recommendations (2022) and the Interim Measures for the Management of Generative AI Services (2023). Key obligations under the AIGC Measures include mandatory visible or implicit marking of AI-generated content at the point of production and distribution, technical traceability requirements enabling content to be traced back to its source AI system, prohibitions on using AI-generated content to spread disinformation or impersonate real individuals without consent, and obligations on internet platform operators to implement detection and labelling capabilities. The measures apply to any organisation or individual generating, publishing, or transmitting AI-generated content through internet information services in China, including foreign entities whose services are accessible to users in China. Enforcement is conducted by the CAC and provincial cyberspace authorities, with penalties including content takedown, service suspension, and fines under the Cybersecurity Law and related instruments.

Key Requirements

• •AI-generated content must be explicitly labelled at the point of generation and at the point of distribution, using both visible labels and embedded technical markers where technically feasible
• •Internet platform operators must implement technical systems capable of detecting AI-generated content uploaded by users and applying required labels
• •Traceability mechanisms must enable regulatory authorities to identify the AI system and entity responsible for generating content when requested
• •It is prohibited to use AI-generated content to fabricate news, impersonate individuals, spread disinformation, or undermine national security and social order
• •Service providers must maintain records of AI-generated content generation logs for a period specified by the authorities, typically no less than six months
• •Foreign-operated services accessible to users in China are subject to the measures regardless of where the service provider is incorporated
• •Non-compliance may result in content removal orders, service suspension, and financial penalties under applicable cybersecurity and internet information service laws

What Your Organization Must Do

• →Audit all AI-generated content workflows, including generative AI tools used for marketing, customer service, and internal communications, to identify which outputs are subject to labelling obligations under the measures
• →Implement technical labelling capabilities including visible watermarks or disclosures and, where technically feasible, embedded provenance metadata such as C2PA-compatible markers for content distributed in China
• →Update terms of service and acceptable use policies for any platform operating in China to explicitly prohibit users from distributing unlabelled AI-generated content
• →Establish content traceability logging systems with retention schedules aligned to regulatory requirements and ensure logs are accessible to CAC upon request
• →Brief legal, product, and communications teams on the prohibition on using AI-generated content for disinformation, ensuring editorial and social media workflows include human review gates for sensitive topics
• →Engage local counsel to assess whether existing generative AI service arrangements with Chinese platform partners satisfy the joint liability implications for platform operators under the measures