Abstract: The development of generative artificial intelligence (AI) is posing profound and subversive challenges to human society. The technological update of GPT-4 also causes risks in underlying algorithm, training data, and intellectual property at the same time. In terms of the underlying algorithm, although there is a risk of algorithm discrimination in GPT-4, it is not necessary to open the algorithm, we should learn from brain-inspired research to promote GPT-4 to AGI; In terms of training data, the massive data behind GPT-4 has a large compliance risk, we should set up a data destruction system, maintain ideological security and explore development options with Chinese characteristics; In terms of intellectual property, GPT-4 brings a series of infringement risks, which has triggered disputes over the attribution of the product, but it cannot constitute a challenge to the qualification of human authors. In order to better deal with the risks in the process of technological development, we should formulate an appropriate regulatory schemes in time, under the theory of meta-regulation, learn from the institutional design in the Digital Services Act of the European Union, combine with existing algorithm governance practice, seek independent innovation in the digital age, help the safe development of the artificial intelligence industry.