Recently, a well-known medical expert was "promoted" in an online video, but in fact, the video was synthesized by AI (artificial intelligence), which has raised public concerns about the risk of AI fraud. Generative artificial intelligence is currently the most popular form of AI technology both domestically and internationally, and is widely welcomed by users due to its powerful content production capabilities and strong interactivity. As a new disruptive technology, generative artificial intelligence brings tremendous benefits and infinite possibilities to economic and social development, but also impacts legal norms, ethics, social security, public governance, and other aspects. One is the endogenous risk caused by the limitations and characteristics of generative artificial intelligence technology itself. The data used by large models contains errors, incompleteness, and inherent human values, which may lead to errors, biases, and discrimination in the output content. Moreover, generative artificial intelligence training relies on large-scale data, which can easily lead to excessive collection of personal and public data, resulting in issues such as infringement. The second is the exogenous risk caused by the illegal and malicious use of some entities. Generative artificial intelligence can become a tool for manufacturing and disseminating false information, and using AI for deep forgery can become a "new weapon" for fraudulent activities. In the context of great power games, generative artificial intelligence may also be used as a tool for cyber attacks. To attach importance to and prevent these risks, it is necessary to strengthen the safety supervision of generative artificial intelligence, establish a solid safety bottom line, promote the improvement of intelligence, and provide guarantees for preventing and resolving related risks and hidden dangers, and continuously promoting industrial development and technological progress. Establish a safety supervision system and conduct full chain supervision. In the preparation stage of generative artificial intelligence, establish standardized mechanisms for data collection and use to avoid data security risks. During the training phase, establish regulatory standards for algorithms, deeply embed legal requirements into the algorithm models of generative artificial intelligence, and eliminate algorithmic bias. In the stage of use, standardize the identification of AI generated content, study and improve relevant laws and regulations on intellectual property protection in the field of generative artificial intelligence, formulate precise and normalized regulatory mechanisms for the abuse and illegal use of generative artificial intelligence, and avoid infringement and disruption of social order. Promote the participation of multiple stakeholders and form a regulatory synergy. Government departments should strengthen top-level design, conduct forward-looking analysis and judgment on the security risk patterns of generative artificial intelligence, actively identify and adapt to changes, improve institutional safeguards, and establish a regulatory system. Consolidate the main responsibility of generative artificial intelligence research and development enterprises, enhance their sense of social responsibility, strengthen industry self-discipline, formulate industry standards, enhance data management and protection, strengthen product safety testing, provide easy to understand product usage guidelines and privacy policies, and develop emergency plans and risk management systems to respond to generative artificial intelligence product safety incidents. Enhance the legal and safety awareness of generative artificial intelligence R&D personnel, conduct AI related legal and regulatory training and safety awareness education, and develop codes of conduct and guidelines for R&D personnel. Improve public AI literacy, strengthen public education on intelligent safety risks, promote and popularize AI related laws and regulations, encourage public participation in the supervision of generative artificial intelligence, establish complaint and reporting mechanisms, and promptly identify and handle generative artificial intelligence safety risk issues. Strengthen technological empowerment and improve the level of safety supervision. At present, China's R&D investment and the number of large models in the field of generative artificial intelligence are constantly increasing, but the quality of large models still needs to be improved. Therefore, it is necessary to increase R&D investment, seize the technological high ground, and ensure the safety and reliability of generative artificial intelligence systems. AI can also serve as an effective technical tool for security supervision, such as using generative artificial intelligence for security risk warning and monitoring, conducting automated compliance checks, identifying and filtering false information, simulating security risk events, helping to develop emergency plans, analyzing the impact and effectiveness of policies, and providing decision support for policy makers. Strengthen international cooperation and promote the formation of international consensus on safety supervision. Generative artificial intelligence governance has become a global technological issue. It is necessary to strengthen international cooperation, establish multilateral dialogue platforms to study and exchange issues related to generative artificial intelligence security regulation, promote the exchange and cooperation of security talents among countries, actively absorb the experience of AI security regulation in various countries, participate in the formulation of international rules for generative artificial intelligence governance and regulation, promote the formation of consensus on generative artificial intelligence security regulation with the concept of a community with a shared future for mankind, jointly address the global security challenges it brings, and promote relevant technologies to serve the development of all mankind. (New Society)