China's Measures for Labelling AI-Generated and Synthetic Content
Issued by
Cyberspace Administration of China, Ministry of Industry and Information Technology, Ministry of Public Security, and National Radio and Television Administration
- September 30, 2026 · Correction — Corrected the issuing agencies (CAC, MIIT, MPS and NRTA), removed an unverifiable link, and merged a duplicate entry. (Cody Maxwell)
- September 30, 2026 · Correction — Corrected the enforcement agencies, the platform duty (label, not block) and the labelling methods (watermarks encouraged, encryption not required) (fact-check finding). (Cody Maxwell)
- October 1, 2026 · Correction — Rewrote the practical steps, newsletter hook, search description, and audience fields to match the entry's corrected content. (Cody Maxwell)
China's labelling measures require AI-generated and synthetic content shown to users in China to carry labels. Visible labels are needed in many cases, and implicit labels must be embedded in file metadata. Platforms must check for labels and add notices to AI-generated content.
Applies To
Overview
Issued on 14 March 2025 by the Cyberspace Administration of China and three other agencies, and backed by the mandatory national standard GB 45438-2025, these Measures establish a comprehensive technical and procedural framework for identifying AI-generated and synthetic content across digital platforms. The regulation covers all forms of AI-generated output including text, images, audio, video, and immersive or virtual reality (VR) content made available to users within China. Key provisions require implicit labels embedded in each file's metadata, and visible labels suited to each format. Watermarking is encouraged but not mandatory, and no encryption is required. The aim is that both automated systems and users can identify AI-origin content. Platforms must check files for implicit labels and add a notice to content that is, or appears to be, AI-generated. They are not required to block unlabelled content. Enforcement falls to the Cyberspace Administration of China and the other issuing agencies within their remits. The Measures take effect in September 2025, giving enterprises a defined compliance window from the March 2025 publication date.
Key Requirements
- •Add visible (explicit) labels to AI-generated text, images, audio, video, and virtual scenes where the measures require them.
- •Embed implicit labels in file metadata, recording generation attributes, the provider's name or code, and a content ID.
- •Digital watermarks for implicit labels are encouraged but not required.
- •Platforms must check metadata for implicit labels and add a notice to content that is, or is suspected to be, AI-generated.
- •The measures and standard GB 45438-2025 have applied since 1 September 2025.
- •Cyberspace, telecommunications, public security, and broadcasting regulators handle violations (Article 13).
What Your Organization Must Do
- →Map every AI-generated text, image, audio, video, and virtual scene your organisation shows to users in China.
- →Add visible labels to that content in the formats the measures require for each media type.
- →Embed implicit labels in file metadata, covering generation attributes, your provider name or code, and a content identifier.
- →Build platform checks that read file metadata and add a notice to content that is, or appears to be, AI-generated.
- →Decide whether to use digital watermarks, which the measures encourage but do not require.
- →Brief legal teams that cyberspace, telecommunications, public security, and broadcasting regulators each handle violations within their remits.
Governance Controls
Operational controls that implement requirements from this regulation.