Chengdu’s Administration for Market Regulation has published the Guidance on Compliance for AI-Generated Advertising. It covers advertising created or synthesized with AI, including text, images, audio, video, virtual scenes, AI avatars, and digital portraites of real people.
The guidance says advertisers and publishers should disclose both the legally required advertising label and the relevant AI-production information, such as AI-assisted optimization, AI-processed visuals, AI virtual scenes, AI-generated content, AI avatars, or digital portraites of real people. The label should help consumers distinguish synthetic content from real experiences and scenes.
Its prohibited list includes false or misleading advertising, illegal or vulgar synthetic content, unauthorized digital portraites, face-swapping that impersonates celebrities or other people, unsupported product experiences claimed by an AI avatar, and content that infringes copyright, trademark, or other third-party rights. AI-generated fake experiences used in reviews, product seeding, or store-visit content are also covered.
The document assigns responsibilities across the chain. Advertisers remain responsible for truthfulness and legality; agencies and publishers should keep intake, review, and archive records; platforms should monitor and handle violations; and AI-generation providers should support required labeling and should not provide services for illegal advertising. The guidance also encourages stricter review for areas such as medicine, finance, education, and products or services for minors.
The guidance applies to commercial advertising activities within Chengdu and should be read together with applicable laws, platform rules, and sector-specific requirements. Read the official Chinese text.
