自动驾驶道德决策的机器研究范式初探  

A Preliminary Study on the Machine Research Paradigm of AutonomousDriving Moral Decision-making

作　　者：宋强[1] 李伦[1] SONG Qiang;LI Lun(Center for Ethical Legal and Social Issues of Big Data and Artificial Intelligence,Department of Philosophy,Dalian University of Technology,Dalian 116024,China)

机构地区：[1]大连理工大学大数据与人工智能伦理法律与社会研究中心、哲学系,大连116024

出　　处：《自然辩证法研究》2024年第7期84-90,98,共8页Studies in Dialectics of Nature

基　　金：国家社会科学基金重大项目“大数据环境下信息价值开发的伦理约束机制研究”(17ZDA023)。

摘　　要：自动驾驶汽车作为一种蕴含“智慧”的新型人工物,其道德决策一直是令人关注的问题。针对这个问题,人们提出了功利主义、义务论、美德论、罗尔斯式以及风险伦理等道德决策方案。然而,这些方案存在一个共同的问题,即采用自动驾驶道德决策的类人研究范式,忽视自动驾驶道德决策的特殊场景。自动驾驶道德决策应采取机器研究范式,这种研究范式下的自动驾驶道德决策应当遵循预防原则和零伤害原则。As a new artificial artifact containing“intelligence”,autonomous driving vehicles’moral decision-making has always been a matter of concern.In response to this problem,people have proposed moral decision-making schemes such as utilitarianism,deontology,virtue theory,Rawlsian and risk ethics.However,these schemes have a common problem,which is to adopt the human-like research paradigm of autonomous driving moral decision-making and ignore the special scenarios of autonomous driving moral decision-making.Autonomous driving moral decision-making should adopt the machine research paradigm.Autonomous driving moral decision-making under this research paradigm should follow the precautionary principle and the zero-harm principle.

关 键 词：自动驾驶 道德决策 类人研究范式 机器研究范式 

分 类 号：N031[自然科学总论—科学技术哲学]