Research on Generative Al Risk Governance and Model Improvement
With the emergence of generative Al technology represented by ChatGPT the original governance model based on traditional Al is gradually unable to adapt to the current new situation.The scenario-based decentralized governance and early prevention supervision concept have forced China to further adjust the governance model and governance concept.This paper focuses on the security risks and governance dilemmas of generative Al,analyzes the shortcomings of China's existing governance models,and discusses intelligent governance models that adapt to the new situation.Literature research and content analysis are used to analyze the limitations of the existing governance system theoretically,and provide a theoretical basis for the new governance model.The results show that rigid norms and supervision alone are not enough to govern the complexity of generative Al,and the new governance model needs to start from four aspects:governance concept,system,norm and mechanism,build a governance system with the participation of multiple subjects,and balance the normative orientation of prudence and inclusiveness.At the same time,it is recommended to establish a whole-process supervision and evaluation model,supported by legal and technical means,to promote the innovation of generative Al governance model,and to achieve a dynamic balance between social development and stability.
万方数据
维普
