从规制到治理:论生成式人工智能法律管理的新范式  

From Regulation to Governance:A new Paradigm of Generative Artificial Intelligence Law Management

作　　者：黄伟文 李昕宇 HUANG Weiwen;LI Xinyu(School of Law,Guangdong University of Finance and Economics,Guangzhou Guangdong 510000,China)

机构地区：[1]广东财经大学法学院,广东广州510000

出　　处：《乐山师范学院学报》2024年第6期94-102,共9页Journal of Leshan Normal University

基　　金：司法部国家法治建设与法学理论研究项目“理性算法信任的法律规制模式研究”(22SFB3005);广东省哲学社会科学规划项目“良法生活视域下算法伦理的法律化研究”(GD23XFX03)。

摘　　要：生成式人工智能(Generative Artificial Intelligence,GAI)的发展突破了人们对技术的认知边界,推动了人类社会的深刻变革。然而,生成式人工智能也给传统法律秩序带来了风险与挑战,而传统的法律规制存在系统性、实效性不足的局限,不能对生成式人工智能进行有效管理。因此,应从法律规制转为法律治理,这就要求在理念上需以创新和安全相协调为原则,以人为中心促进人机关系的合作共赢;在法律治理实践中建立分级治理架构,完善多主体审查体系并强化应用合规的激励制度,促使生成式人工智能的发展符合法律与伦理的要求。The development of Generative Artificial Intelligence(GPT)has broken through the boundaries of people's cognition towards technology,driving profound changes in human society.However,generative artificial intelligence also brings risks and challenges to the traditional legal order.Moreover,due to limitations in terms of systemic and practical effectiveness,the traditional legal regulations cannot effectively manage generative artificial intelligence.Therefore,we should shift from legal regulation to legal governance,which requires that innovation and security should be coordinated as the principle in concept,with a people-centered approach to promote a win-win relationship between humans and technology.In the practice of legal governance,it is necessary to establish a hierarchical governance structure,improve the multi-agent review system,and promote the incentive system for compliance application.Thus,the development of generative artificial intelligence meets the requirements of law and ethics.

关 键 词：生成式人工智能 法律风险 法律治理 

分 类 号：D922.17[政治法律—宪法学与行政法学] D922.16[政治法律—法学]