Abstract: Improving the credible governance system of artificial intelligence is necessary to promote technological innovation and a legal guarantee to comprehensively respond to technological alienation and promote the development of new quality productive forces. Uncontrollable artificial intelligence has triggered a serious crisis of confidence in its application and even threatened social stability and national security. To balance development and security, the governance consensus of "credible development" should be condensed, and the flexibility and effectiveness of governance should be improved by following the theory of life cycle and technical governance. On the one hand, improves the legal compliance guidance framework before, during, and after the event, and reduces regulatory costs while improving regulatory deterrence; on the other hand, it is necessary to innovate the technology governance system, introduce regulatory technologies such as private cloud computing, supervision fine-tuning, and SAGE analysis, and improve the human-machine collaborative framework to fully implement data, algorithms, and content credible regulatory requirements. To improve the artificial intelligence safety supervision system, prevent the risk of alienation, and create a "mutually beneficial symbiosis" man-machine relationship.

摘要： 完善人工智能可信治理体系不仅是促进技术创新之必需，也是全面回应技术异化并助推新质生产力发展的法治保障。不可控但是快速迭代的人工智能在应用中引发严重信任危机，乃至会威胁社会稳定和国家安全。为兼顾发展与安全目标，应凝聚起“可信发展”这一治理共识，并遵循生命周期和技术治理理论，提升治理的灵活性和有效性。一方面，健全事前、事中以及事后的法律合规指引框架，在降低监管成本同时提升监管威慑性；另一方面，需创新技术治理体系，引入私有云计算、监督微调以及SAGE分析等监管科技，健全人机协作框架，以全面落实数据、算法以及内容可信监管要求。由此健全人工智能安全监管制度，防范异化风险并营造“互利共生”人机关系。