面向技术本身的人工智能伦理框架——以自动驾驶系统为例  

A Framework of Technology-oriented Ethics for Artificial Intelligence:On Self-driving Systems

作　　者：潘恩荣[1,2] 杨嘉帆 PAN Enrong;YANG Jiafan(School of Marxism,Zhejiang University,Zhejiang,Hangzhou,310028;International Design Institute,Zhejiang University,Zhejiang,Hangzhou,310028)

机构地区：[1]浙江大学马克思主义学院,浙江杭州310028 [2]浙江大学国际设计研究院,浙江杭州310028

出　　处：《自然辩证法通讯》2020年第3期33-39,共7页Journal of Dialectics of Nature

基　　金：国家社会科学基金重大项目“中国工程实践的伦理形态学研究”(项目编号:15ZDB015)。

摘　　要：人工智能产业界在解决伦理问题时发现,技术专家无能为力。当转而寻求伦理专家给出方案,如自动驾驶在电车难题情景中撞向谁的伦理决策方案时,伦理专家发现自身难以承受如此责任。从现代技术哲学经验转向视角来看,这是技术和伦理相互独立甚至对立引发的后果。理想的治理方式是将人工智能技术与伦理融合成一种面向技术本身的人工智能伦理框架。基于“设计情景-使用情景”,结合人工智能的两种技术设计思路“人工的智能”(AI)和“智能增强”(IA),这样的伦理框架能够解决自动驾驶面临的问责难题。The artificial intelligence industry has found out that technical experts do not have much to offer when dealing with ethical issues.Meanwhile,when an ethical decision needs to be made for something like The Trolley Problem,experts in ethics also find it too heavy a burden to shoulder.From the perspective of Empirical Turn of Modern Philosophy of Technology,this is the result of technology and ethics being mutually independent and even opposite to each other.The best solution would be to integrate artificial intelligence technology and ethics so as to come up with an ethics framework for artificial intelligence that is technology-oriented.Based on combining the two design ideas of artificial intelligence technologies,Artificial Intelligence(AI)and Intelligence Augmentation(IA),in the context of design and use,such an ethics framework would bring into existence an effective accountability mechanism for self-driving systems.

关 键 词：自动驾驶 面向技术本身 经验转向 伦理决策 智能增强 

分 类 号：N0[自然科学总论—科学技术哲学]