deep synthesis
In plain wordsThe Chinese legal term for AI that creates or alters text, images, voices, video or virtual scenes — the technology behind deepfakes — and for the rules that govern it.
In Chinese regulation, the use of generative algorithms, from deep learningdeep learning, n.: Machine learning done with neural networks built from many stacked layers: raw data such as pixels, sound or text goes in at one end, and each layer turns what it receives into something more abstract. to virtual-reality techniques, to create or alter content of any kind (writing, voices and other sound, pictures, video, virtual scenes), as in face swapping, voice cloning, automated writing and 3D reconstruction. The Provisions on the Administration of Deep Synthesis Internet Information Services, in force since January 10, 2023, require providers to verify users' real identities; obtain consent before editing a person's face or voice; review inputs and outputs; prominently label content that could confuse or mislead the public, such as synthetic faces and voices, chatbot dialogue and immersive scenes; prohibit the removal or concealment of labels; keep logs; complete algorithm filingalgorithm filing, n.: China's requirement that companies register certain algorithms — such as recommendation feeds and chatbots — with the internet regulator, describing what they do and how their risks are controlled.; and assess the security of tools that edit biometric features. Using deep synthesis to produce or spread fake news is forbidden. The term covers what English speakers call deepfakesdeepfake, n.: A fake but realistic video, photo or voice recording made with AI that puts words in a real person's mouth or shows them in a scene that never happened. and synthetic mediasynthetic media, n.: Images, video, audio or text created or substantially altered by AI rather than recorded or written directly by people, whether for art, entertainment, business or deception. while avoiding the pejorative 'fake'.
Because the app swaps faces, it counts as a deep synthesis service and must label its output.
Origin
Used in Chinese industry and policy writing by the early 2020s as a neutral alternative to 'deepfake'; codified in rules issued on November 25, 2022 by the Cyberspace Administration of China, the Ministry of Industry and Information Technology and the Ministry of Public Security.
Usage note
Its labeling duties were elaborated by China's 2025 AI-generated content labelingAI-generated content labeling, n.: Marking text, images, audio or video as made by AI — with a visible notice, hidden data in the file, or both — so that people and platforms can tell synthetic content from real. measures and mandatory national standard.
Related
- deepfakedeepfake, n.: A fake but realistic video, photo or voice recording made with AI that puts words in a real person's mouth or shows them in a scene that never happened.
- synthetic mediasynthetic media, n.: Images, video, audio or text created or substantially altered by AI rather than recorded or written directly by people, whether for art, entertainment, business or deception.
- AI-generated content labelingAI-generated content labeling, n.: Marking text, images, audio or video as made by AI — with a visible notice, hidden data in the file, or both — so that people and platforms can tell synthetic content from real.
- algorithm filingalgorithm filing, n.: China's requirement that companies register certain algorithms — such as recommendation feeds and chatbots — with the internet regulator, describing what they do and how their risks are controlled.
- Interim Measures for the Management of Generative AI ServicesInterim Measures for the Management of Generative AI Services, prop. n.: China's 2023 rules for chatbots and other AI that generate text, images, audio or video for the public, requiring lawful training data, content controls, labeling and, for many services, government approval before launch.
- Cyberspace Administration of ChinaCyberspace Administration of China, prop. n.: China's top internet authority, in charge of online censorship as well as regulation, which also writes and enforces most of the country's AI rules, such as the registration of chatbots and recommendation algorithms.
Sources
- DLA Piper. Artificial Intelligence in China (regulatory tracker).
Defined in law or standards
Provisions on the Administration of Deep Synthesis Internet Information Services (CAC, MIIT and MPS, 2022)
Mentioned in
AI-generated content labelingAI-generated content labeling, n.: Marking text, images, audio or video as made by AI — with a visible notice, hidden data in the file, or both — so that people and platforms can tell synthetic content from real., Cyberspace Administration of ChinaCyberspace Administration of China, prop. n.: China's top internet authority, in charge of online censorship as well as regulation, which also writes and enforces most of the country's AI rules, such as the registration of chatbots and recommendation algorithms., Interim Measures for the Management of Generative AI ServicesInterim Measures for the Management of Generative AI Services, prop. n.: China's 2023 rules for chatbots and other AI that generate text, images, audio or video for the public, requiring lawful training data, content controls, labeling and, for many services, government approval before launch., synthetic mediasynthetic media, n.: Images, video, audio or text created or substantially altered by AI rather than recorded or written directly by people, whether for art, entertainment, business or deception., algorithm filingalgorithm filing, n.: China's requirement that companies register certain algorithms — such as recommendation feeds and chatbots — with the internet regulator, describing what they do and how their risks are controlled.