Mode Selection and System Expansion of China's Artificial Intelligence Legislation—Also on the Domain-Integrated Legislative Model
Since the 20th century,AI has been continuously promoting the transformation and upgrading of traditional industrial society to a digital society.In this process,while AI brings opportunities,it also triggers many risks such as privacy leaks and algorithmic discrimination,making AI governance gradually become the new focus of global intelligent competition.In April 2021,the European Commission issued the"Artificial Intelligence Act",and continued to advance it over several years.At the same time,the United States also actively tried to govern AI through legislation and maximized the support for innovative development.In October 2023,Biden issued the"Executive Order on the Safe,Secure,and Trustworthy Development and Use of Artificial Intelligence",proposing eight objectives such as AI safety and security,protecting the privacy of American people,and promoting innovation and competition.As for China,in June 2023,the State Council included the AI Law in the"2023 Annual Legislative Plan".During the Two Sessions in 2024,several members of the CPPCC suggested that the AI legislation should be promoted as soon as possible.However,existing academic research rarely discusses AI legislation specifically,and more often bases discussions on data and algorithms,such asdata rights confirmation,algorithmic discrimination and interpretability.The above research is certainly beneficial for AI governance,but the governance of intelligent elements cannot replace AI governance.On the whole,most of the relevant research on AI governance is from the perspectives of technology ethics,national policies,and advocacy principles,which makes the content present departmental law colors and concrete characteristics,and it is easy to produce regulatory competition or regulatory blind spots in practice.In fact,it is because of the absence of macro-level AI legislation that the existing normative content overlaps and the order of application is chaotic.Compared with previous literature,this article believes that the clarification of the legislative model is the primary issue,so it chooses a new research topic.This article mainly uses comparative research methods to investigate the practice of AI legislation in the international community,and summarizes two types:the unified vertical legislative model and the dispersed grid legislative model.The unified vertical legislative model refers to the state formulating a special AI bill or code,making unified provisions on intelligent definitions,regulatory agencies,etc.,to ensure the internal consistency of the normative system and carry out vertical and systematic governance of AI.Representative countries are the European Union and Canada.The dispersed grid legislative model refers to the states which are not formulating a unified regulatory bill,but forming individual norms of different grid nodes by distinguishing intelligent types and application scenarios,and carrying out flexible and targeted governance of AI.Representative countries are the United States and the United Kingdom.After further analysis,it is found that both models are adapted to the specific situations of certain countries or regions,and can fully exert their advantages.However,China's industrial status,strategic policies,and political system have special features,and blind copying may be counterproductive.Based on the research results of the academic community,this article uses the Law 3.0 theory as a legislative guide,and proposes that China should explore the domain-integrated legislative model.This model emphasizes that AI legislation is a domain law,and the content should include risk regulation,industry promotion,and technical standards.In terms of form,it should be based on the basic law,with industry promotion law and technical standard law as branches.Compared with the existing two models,it is more conducive to achieving the triple integration of legislative purpose,AI orientation,and rule of law system engineering.In addition,considering that the promotion of the"Artificial Intelligence Law"is an urgent task,this article has made unified definitions,absorbed referrals,and reshaped the system for basic provisions,basic principles,basic systems,and the responsibility and obligation system,which will help the next step of legislative work and theoretical research to some extent.
AI legislationintelligent governanceunified vertical legislationdispersed grid legislationdomain-integrated legis-lation
国家科技期刊平台
NSTL
万方数据
