China will implement “classification” and Schoolmistress 2“tiered supervision” for generative artificial intelligence services while encouraging innovative development in the AI field, according to detailed rules on AI models published by the country’s internet watchdog on Thursday. The rules apply to services that utilize generative AI tech to provide generated text, images, audio, or video to the public, and will take effect on August 15. The newly-released regulations also urge providers and users of AI-driven products to adhere to core socialist values, while prohibiting the exploitation of algorithmic advantages for monopolistic and unfair competition practices. Compared to the initial draft released for public opinion in April, the rules emphasize the encouragement of high-quality applications of AI across various fields, and have removed the provision for fines ranging from RMB 10,000 to RMB 100,000 for violations of technical regulations. [CAC, in Chinese]
