China's generative AI rules carved out film and TV production

What was announced or recorded

On 13 July 2023 the Cyberspace Administration of China (CAC) published Order No. 15, the Interim Measures for the Administration of Generative Artificial Intelligence Services, jointly issued with the national development and reform, education, science and technology, industry and information technology, and public security authorities and, notably for micro-drama regulation, the National Radio and Television Administration (NRTA). The measures were signed on 10 July 2023 and took effect on 15 August 2023.

What the documents establish

The final text applies to any service generating text, image, audio or video content for the public within mainland China using generative AI, but its Article 2 adds a carve-out: where the state has separate rules governing press and publishing, film and television production, or literary and artistic creation using generative AI, those separate rules apply instead. Comparing that text against a DigiChina translation of the April 2023 draft for comment, used here as a labelled English-language reference rather than as evidence of the final rule's exact wording, shows the carve-out was new: the draft's own Article 2 applied to any generative AI product or service offered to the public in China, with no such exception and no exemption for non-public-facing research either, both added by the final version. The final measures separately require providers to label AI-generated images and video under China's Deep Synthesis Provisions, disclose their service's intended users and uses, and take effective measures against minors over-relying on or becoming addicted to the service.

Announced versus delivered

The carve-out means an AI-made micro-drama does not exit China's regulatory system by virtue of being AI-generated; it stays inside whichever system already governs micro-dramas as film and television production, NRTA's licensing and tiered filing rules, with these measures' labelling and minors provisions layered on top rather than substituted in. The text does not itself state how the two regimes interact case by case, and no fully AI-generated drama's registration path is described here.

What to watch

This is an editorial test, not a stated one: whether a micro-drama that is substantially or fully AI-generated carries the deep-synthesis label these measures require, in addition to whatever licence or registration number NRTA's own rules require, since the measures state that both apply rather than one displacing the other.

• Does a promoted 'AI-generated' or 'AIGC' micro-drama carry a visible deep-synthesis label, the disclosure Article 12 requires for generated images and video?
• Is the drama distributed under NRTA's film-and-television licensing route, the separate regime Article 2 preserves, or is a provider treating AI generation as exempting it from that route?
• Does a platform hosting AI-made dramas publish the target-audience and minors-protection disclosures Article 10 requires, separately from any content rating the drama itself carries?

The measures are best read as adding a second, technology-specific layer of obligation on top of micro-drama regulation that already existed, not as a new gate that fully generated content passes through instead of the old one.