The “Regulation on the Application of Deep Synthesis in Internet Information Services” will be implemented on Jan. 10 2023
By using AI technology, you may edit existing works in a depth synthesis manner, such as to replace the actor’s face or the original speech, or to cut partial of the works. Those editing brings fun to readers, but also risks of infringement to editors. The risks of infringement mainly include the infringement of the copyright of the work itself, the infringement of the reputation of the synthesized object, and even some people may constitute the crime of fraud by using the edited works. Therefore, the Cyberspace Administration of China, the Ministry of Industry and Information Technology and the Ministry of Public Security jointly released the “Regulation on the Application of Deep Synthesis in Internet Information Services” (hereinafter referred to as the “Regulation”) on Nov. 25, 2022, which will be implemented on Jan. 10, 2023. The following is an introduction to the main contents of the Regulations.
- To clarify the definition of relevant terms of deep synthesis technology as follows:
1) The deep synthesis technology refers to the technology using deep learning, virtual reality and other generative synthesis algorithms to produce network information such as text, images, audio, video, and virtual scenes, and etc. The network information mainly including the following categories:
a) Technologies to generate or edit text content, such as text generation, text style conversion, Q&A dialogue and so on;
b) Technologies to generate or edit speech content, such as text to speech, speech conversion, speech attribute editing and so on;
c) Technologies to generate or edit non-speech content, such as music generation, scene sound editing and so on;
d) Technologies to generate or edit the biometrics in images and video content, such as face generation, face replacement, character attribute editing, face manipulation, pose manipulation and so on;
e) Technologies to generate or edit the non-biometrics in images and video content, such as image generation, image enhancement, image restoration and so on; and
f) Technologies to generate or edit digital characters and virtual scene, such as 3D reconstruction, digital simulation and so on.
2) The deep synthesis service involves three types of targeted entities as follows,
a) The deep synthesis services provider refers to entities or individuals providing deep composition services.
b) The technology supporter for deep synthesis services refers to entities or individuals providing technical support for deep synthesis service.
c) The deep synthesis services user refers to entities or individuals that use deep synthesis services to produce, copy, publish and disseminate information.
3) Training data refers to the labeled or benchmark data set used to train machine learning models.
4) Immersive virtual scene refers to a virtual scene generated or edited by depth synthesis technology, which can be experienced or interacted by participants and has a high degree of realism.
- The “Regulation” mainly puts forward requirements for deep synthesis service providers
The requirements for deep synthesis service providers as stipulated in the “Regulation” are in line with the requirements as stipulated in the “Network Security Law”, the “Data Security Law” and the “Personal Information Protection Law”. For the majority of deep synthesis services users, the two prohibited acts listed in Article 6 of the “Regulation” deserves special attention. The two prohibited acts are, 1) users should not use the deep synthesis service to produce, copy, publish, and disseminate information prohibited by laws and administrative regulations; and 2) users should not use the deep synthesis service to engage in activities prohibited by laws and administrative regulations that endanger national security and interests, damage the national image, infringe on social and public interests, disrupt economic and social order, and infringe on the legitimate rights and interests of others.