欧美人工智能治理模式比较研究

A comparative study of European and American AI governance models

当前,AI治理逐渐走向法律治理时代,且呈现明显的“国别差异”特征。最有代表性的AI治理模式,分别是欧盟的以伦理优先为导向的集中式治理模式和美国以创新优先为导向的分散式治理模式。研究美、欧AI治理模式的内在形成逻辑、特征及其所面临的困境,对于反思、构建符合本国国情的AI治理体系大有裨益。文章从治理理念、治理架构、治理主体、治理强度、治理困境五个方面比较分析了美国、欧盟AI治理体系的差异。在此基础上,探讨了我国人工智能治理体系构建过程所应注意的普遍问题与原则,如重视路径依赖与模式移植间的张力、平衡“发展”与“伦理”间的关系、独立自主与国际融入间的协调、国家规制与多元参与间的协同等。

At present,AI governance is gradually moving towards the era of legal governance,with obvious"national differences"characteristics,a series of factors such as existing AI technology reserves,business ecology,cultural and legal traditions,social management systems,human resources,and governance stages often shape the AI governance model of a country or region,can result in different AI ethical governance concepts and paths.The most representative AI governance models presently are the EU's centralized governance model guided by ethics priority and the US's decentralized governance model guided by innovation priority.The study of the internal formation logic and characteristics of different AI governance models is of great benefit to the reflection and construction of AI governance systems in line with national conditions.The article compares and analyzes the differences of AI governance system between the United States and the European Union from the following five aspects:(1)The perspective of governance idea comparison.One of the prominent features of the EU AI governance model is"ethical priority",which means strengthening the EU's influence and leadership in the world AI field through the construction of ethical principles and legal systems,so to make up for its shortcomings in AI innovation and application.In comparison,the United States adheres to the principle of"innovation priority orientation",which strives to maintain its AI innovation capability,and holds a balance between"technological innovation"and"application acceptability";(2)The perspective of AI governance structure comparison.The EU has obvious characteristics of unified governance,it has constructed a set of vertical governance systems;In contrast,the United States as a whole presents a"dual grid"type of multi-headed and decentralized governance characteristics,with independent legislation in each state;(3)The perspective of AI ethical governance entities comparison.American enterprises have strong ethical autonomy capabilities and stronger bottom-up governance signs,but for the EU,AI ethical governance is more dependent on the government,presenting a vertical governance feature from top to bottom,and the ability of enterprise autonomy is relatively weak;(4)The perspective of governance intensity comparison.the EU model has obvious"strict"governance characteristics,while the US model exhibits"enterprise friendly"and"weak governance"characteristics,giving enterprises more room for fault tolerance;(5)The perspective of governance dilemma comparison.An important problem facing the EU is the entanglement of governance ideas,that is,how to balance the relationship between technological innovation and strict governance,as EU AI governance shows strong ethical constraints characteristics,but this mode also drags down the overall pace of its innovation progress;For the United States,due to its federal system,there are various cultural,legal,and interest conflicts between the federal government and states,as well as between states,so this special political system makes it face the dilemma of rules fragmentation.Based on the above analysis,the paper discusses the common issues and principles that should be paid attention to during the construction of China's artificial intelligence governance system,such as:(1)To maintain an appropriate tension between path dependence and model transplantation,it is not reasonable to directly embed the AI governance concepts and methods of other countries into our own country;(2)To balance the relationship between"development"and"ethics",we should notice that ethics is not the only criterion that should be considered,other social dimensions such as people's livelihood and economy also should be put into attention;(3)To coordinate the relationship between independence and international integration,on the one hand,we should adhere to the principle of"openness"and actively integrate into the globalization process of AI governance,on the other hand,we should not completely passively accept the governance norms of other countries,but actively to build our own standards and seek more international support;(4)To emphasize the synergy between national regulation and multiple participation,we should encourage multiple actors to take part in the governance process,and combine the top-down and bottom-up models altogether.