人工智能政策:入门与路线图  

Artificial Intelligence Policy:Introduction and Roadmap

作　　者：雷恩.卡罗 郑志峰[2,3,4] Ryan Calo (USA);Translated by ZHENG Zhi-feng

机构地区：[1]西华盛顿大学法学院 [2]西南政法大学 [3]贵州省社科院 [4]西南政法大学民商法学院

出　　处：《求是学刊》2019年第2期25-44,共20页Seeking Truth

基　　金：司法部青年项目"自动驾驶汽车的私法挑战与应对研究"(17SFB3031);教育部2019年度青年项目"民法典编纂视野下人工智能侵权责任研究";西南政法大学人工智能法律研究院2018年科研创新重点项目"自动驾驶汽车的法律规制研究"(2018-RGZN-JS-ZD-03)

摘　　要：新一轮人工智能热潮至少有两点独特之处:一是得益于计算能力和训练数据的巨大增长,机器学习取得实质性突破,促使人工智能的大规模应用成为可能;二是决策者终于给予了密切的关注。当前,人工智能引发了一系列严峻的政策挑战,包括公正与平等、武力使用、安全与认证、隐私和权力、税收和失业以及机构配置与专业知识、投资和采购、消除归责的障碍、人工智能的心理模型等跨领域问题。人工智能末日论反映了人类对于人工智能等拟人化科技的特殊恐惧,在可预见的未来并不会真实发生。相反,对人工智能末日论投入过多的关注和资源,可能会分散决策者对于人工智能更直接的危害和挑战的注意力,进而阻碍有关人工智能对当前社会影响的研究。There are at least two unique features in the new wave of artificial intelligence(AI)boom.First,machine learning,having benefited from the huge growth of computing power and training data,has achieved substantial breakthroughs,which makes large-scale application of AI possible.Second,policymakers have finally paid close attention to it.Currently,AI has triggered a series of severe policy challenges,including interdisciplinary issues such as justice and equality,use of force,security and certification,privacy and power,taxation and unemployment,as well as institutional allocation and expertise,investment and procurement,barriers to eliminate accountability,and the psychological model of AI.The doomsday prediction of AI reflects the special fears of human beings for anthropomorphic technologies such as AI,which will not really happen in the foreseeable future.On the contrary,too much attention and resources devoted to the doomsday prediction of AI may distract decision-makers from the more direct harms and challenges of AI,thus hindering the research on the impact of AI on current society.

关 键 词：人工智能 政策挑战 机器学习 人工智能末日论 

分 类 号：TP18[自动化与计算机技术—控制理论与控制工程]