Measures for the Administration of Labelling of AI-Generated and Synthetic Content

Summary

China's Cyberspace Administration published the Measures for the Administration of Labelling of AI-Generated and Synthetic Content on 14 March 2025. The measures took effect on 1 September 2025 and require both visible labels and embedded watermarks on all AI-generated text, image, audio, video and virtual-scene content. Their scope is broader than any other national AI labelling law, explicitly including immersive environments. The measures apply to any provider serving users in China, making them relevant to international platforms and students creating content for Chinese audiences.

Implications for animation education

Animation educators teaching publication and distribution workflows should include these measures in comparative disclosure curriculum alongside California SB 942 and the EU AI Act. The measures apply to any provider serving users in China, which includes international streaming platforms and digital publishers with Chinese audiences. For animation programs with students creating work for Chinese platforms or markets, the embedded watermarking and visible labelling requirements apply from September 2025. The inclusion of virtual scenes (immersive environments) in the scope of the measures is significant for extended-reality and interactive projects. Speculative implication: students using covered AI tools to produce content distributed in China may be required to carry dual disclosure marks from both Chinese and California systems, requiring understanding of both regimes.

Related items

• California SB 942: California AI Transparency Act
• C2PA Content Credentials: Specification and 2024 Adoption Milestones

Source

Source: Cyberspace Administration of China (CAC) ↗ (Policy Text)

Primary document: cac.gov.cn/2025-03/14/c_1743654684782215.htm ↗