摘要
信息社会中传统算法正转向自主(学习)算法,决策主体也由人类主体优位转向算法主体优位。随之出现如何理解和互动于经算法生成之世界的困惑,形成对算法设计、算法(半)自主操作和人之操作之间伦理意蕴的局限性认知,影响我们对“数据化世界”的把握。在此,通过勾勒算法伦理的认识论和规范性两个领域结构图,厘清算法中证据确证性、决策透明度、价值偏见、不公正结果、自主性和隐私性挑战等六个维度伦理争论问题。最后,也指出相应算法伦理争论可能分化和演进的方向。
There is turn from traditional algorithms to autonomous(learning)algorithms in the information society,and the decision-making agents are also shifted from the human-dominated to the algorithm-dominated.In this way,there is some confusion about how to understand and interact with the transformative world which is generated by the algorithm.A limited understanding of the ethical implication among the designs of algorithm,the algorithmic(semi-)autonomous operations and the human operations,inevitably affects our good grasp of the“digitization world”.By providing the prescriptive map of two fields concerning the epistemology and normativeness of algorithmic ethics,this paper clarifies six ethical debates including inconclusive evidence,inscrutable evidence,value bias,unfair outcomes,challenges of autonomy and informational privacy.Finally,further research on the ethical debates of algorithm may be identified.
作者
汪德飞
WANG De-fei(College of Marxism,Huzhou University,Huzhou Zhejiang 313000,China)
出处
《科学技术哲学研究》
CSSCI
北大核心
2021年第4期59-65,共7页
Studies in Philosophy of Science and Technology
关键词
数据化世界
传统算法伦理
自主算法伦理
伦理学争论
digitization world
traditional algorithmic ethics
autonomous(learning)algorithmic ethics
ethics debate
