摘要
算法行政作为一种新型的政府购买服务,其隐秘性、自主决策性、多元主体间行为的不可分性给传统责任框架造成了巨大挑战,可能产生责任鸿沟与治理缝隙。现有治理体系多采取主体行为协同、责任分离的逻辑,容易导致责任逃逸与代理洗白等问题,理想型意义上的穿透式责任面临集合行为与科层制难以穿透的实施困境,因此提出分布式责任对二者进行调和。分布式责任有较为严格的前置适用条件,可视为疑难案件的有效解决方案。分布式责任意味着从中心式的控制风格向全流程参与式问责的转换,在无法明确主观故意的算法致害中实现了对多主体间共同责任的简化。
As a new type of government purchase service,algorithmic administration has posed a great challenge to the traditional responsibility framework due to its secrecy,partly independent decisionmaking and indivisibility of behaviors.Algorithmic administration may lead to responsibility gap and governance gap.The existing governance system mostly adopts the logic of separated responsibility,which cannot penetrate the multiple entrustment relationship between subjects.The separation of responsibilities in practice leads to the problems of responsibility escape and agency washing,while the ideal type of penetrating responsibility faces the implementation dilemma that the collective responsibility and the bureaucracy are difficult to penetrate,so the distributed responsibility is proposed for reconciliation.Distributed responsibility has strict pre-application conditions and can be regarded as an effective solution to difficult cases.Distributed responsibility means the transformation from a centralized control style to a full-process accountability,and the simplification of the joint responsibility among multiple agents in the algorithmic harm where subjective intent cannot be clearly defined.
作者
肖梦黎
Xiao Mengli(Law School of East China University of Science and Technology;Legal Sociology Research Center of East China University of Science and Technology)
出处
《国家检察官学院学报》
CSSCI
北大核心
2023年第2期42-56,共15页
Journal of National Prosecutors College
基金
司法部法治建设与法学理论研究青年项目“公共算法决策的法律责任体系研究”(22SFB4001)的阶段性成果
关键词
数字政府
算法行政
责任体系
穿透式责任
分布式责任
Digital Government
Algorithm Administration
Responsibility System
Penetrating Responsibility
Distributed Responsibility
作者简介
肖梦黎,华东理工大学法学院讲师、法律社会学研究中心副主任。
