Since 15th August, the PRC government has introduced 24 new rules to regulate generative artificial intelligence.

This document is the result of long discussions between the government and business. As a result, the most severe provisions have been removed - for example, the provision for fines of up to 100,000 yuan.
So, what is proposed in these 24 rules?

Firstly, all services and algorithms based on artificial intelligence, which were created or sold in China, now must be compulsory registered with regulatory authorities. There will be seven of these authorities, including the Chinese Ministry of Education, the Ministry of Science and Technology, the National Development and Reform Commission, and, of course, the country’s main authority in the field of Internet regulation – the Cyberspace Administration of China (CAC).

Further, any content created by AI will be marked, so as not to confuse the average user. This rule was issued to prevent the creation of “deepfakes”. Also, companies introducing an AI model must use “legitimate data” to train their models and disclose that data to regulators as required. If not, fines can reach 100,000 yuan.

In addition, the creators of generative AI content must check its safety for users before commercial implementation.

Finally, all Chinese companies operating in the field of AI technologies must provide (for the consideration of the PRC authorities) a clear and transparent mechanism for handling complaints about services or content created by AI.

Thus, the Chinese government endeavors to support the development of AI and, at the same time, maintain the ability to control the information environment in the country.