随着人类社会智能化程度不断提升,人工智能的普及、更新与迭代在极大提高其运行效率的同时,也带来了新的监管困境与治理风险——如何保证人工智能系统的应用是可追溯的、负责任的、公平的,目前我国学术界对这一问题的探讨仍显不足.以公共责任为导向的美国《人工智能监管框架》,在推动各主体在治理中更负责任地使用人工智能系统,确保人工智能治理符合公共责任目标等方面发挥了积极作用.本文通过剖析《框架》的基本理念与关键做法,从构建治理框架,统一基本监管过程和治理结构;完善数字人才队伍,提升实体职业素养和创新能力;编写说明文档,增强系统的解释性和可预测性;推进分级治理,明晰包容治理和审慎监管边界四个方面,论述实现负责任的人工智能治理的中国方案.
With the continuous improvement of the intelligence level of human society,the popularization,updating and iteration of artificial intelligence greatly improve its operation efficiency,but also bring new regulatory dilemmas and governance risks—how to ensure that the application of artificial intelligence system is traceable,responsible and fair,the current discussion on this issue in Chinese academia is still insufficient.The"Artificial Intelligence Regulatory Framework"oriented by public liability in U.S.A has played a positive role in promoting the more responsible use of AI systems in governance by various subjects and ensuring that AI governance is in line with the goal of public responsibility.This paper analyzes the basic concepts and key practices of the Framework,and discusses China's four aspects of realizing responsible AI governance,namely,constructing a governance framework to unify the basic regulatory process and governance structure;perfecting the digital talent team to enhance the professionalism and innovation capability of entities;preparing explanatory documents to enhance the interpretability and predictability of the system;and promoting hierarchical governance to make the boundaries of inclusive governance and prudential regulation clear.