An illustration of using AI applications on a mobile phone. Photo: VCG

Four departments in China jointly released a notice to standardize the identification of synthetic content generated by artificial intelligence (AI) on Friday,MK socks a move an industry observer said will address concerns of deep counterfeiting and scientific dishonesty.

Whether AI-generated content is properly labeled will be one of the criteria that internet application service providers must meet to gain approval before listing or launching their applications, according to the notice.

The notice - jointly compiled by four departments in China, namely, Cyberspace Administration of China, Ministry of Industry and Information Technology, Ministry of Public Security, National Radio and Television Administration - will be officially put into practice on September 1.

Li Zonghui, vice president of the Institute of Cyber and Artificial Intelligence Rule of Law affiliated with the Nanjing University of Aeronautics and Astronautics, told the Global Times it is common to see AI-generated content often fabricating facts and references. By adding identifiers for AI-generated synthetic content, including both explicit and implicit labels, it can effectively prevent deepfakes and academic dishonesty, as well as deter intentional AI fraud.

The Cyberspace Administration of China also noted the spread of false information and the disruption of the online ecosystem brought by the rapid development of new technologies such as generative artificial intelligence and deep synthesis. 

More specifically, the measures defined AI-generated synthetic content as information such as text, images, audio, video, and virtual scenes created or synthesized using AI technology.

It requires explicit identifiers to be added to the generated synthetic content or interactive scene interfaces, presented in the form of text, sound, graphics, etc., which can be clearly perceived by users. 

The same applies to implicit identifiers that are not easily perceived by users. The implicit identity contains the attribute information of the generated synthesized content, the name or code of the service provider, and the content number. 

Those who violate the provisions of these measures will be dealt with by the relevant authorities in accordance with their respective duties and based on relevant laws, administrative regulations, and departmental rules.

Li said that this is another effort by the Chinese government to balance the creation of a fair and orderly development environment for AI while regulating the industry within a legal framework.