In April 2023, China published a discussion draft of the “Regulations for the Administration of Generative Artificial Intelligence Services.” The aim of the document is to ensure content security, compliant operation, and protection of user rights when using generative AI.
The regulations emphasize the legality and ethical obligations of generative AI services. During the content generation process, generative AI should reflect Chinese core socialist values and must not contain illegal information or false or discriminatory content to protect national interests and public order. In addition, the regulations call for the protection of users’ personal information and privacy. The origin of the training data and the rules for labeling generative AI products are also observed.
In case of detected violations or user-reported content that does not meet the requirements, content filtering measures are to be taken in the short term and the model and training are to be optimized within three months. This is to prevent repeated generation. It is also required that providers of generative AI services must label the generated texts, images, videos, etc.
With the publication of the discussion draft, China underlines its ambition to set clear standards for the development of generative AI. It emphasizes that ethics and social responsibility should be observed in the use of generative AI in order to unleash the full innovation potential of AI within a legal and ethical framework.
Picture> AI Image Creator from Microsoft Bing