China mandates algorithmic registration and usage limits for AI

China enacted AI regulations approximately five years before the EU's AI Act, including an algorithm registration requirement where models must be checked against Communist Party values and societal norms before market release. The regulations stipulate that only 10 percent of training data may come from illegal sources. To prevent dependency, the government limits children's social media usage to four hours daily and requires chatbots to remind users every two hours that they are interacting with an AI. These measures aim to balance the promotion of AI adoption with the mitigation of risks related to critical text generation and social dependency, particularly given China's low birth rate. The speaker views China's AI regulatory framework as smart and well-thought-out, suggesting that the concept of model governance and registration is a positive idea that could be adopted globally, even if the specific values differ from Western norms.