算法决策对民主生活的挑战及其应对  

The Challenge of the Algorithmic Decision⁃making to Democratic Life and Its Countermeasures

作　　者：周恒[1] ZHOU Heng(Law School of Zhengzhou University,Zhengzhou,Henan 450001,China)

机构地区：[1]郑州大学法学院,河南郑州450001

出　　处：《时代法学》2023年第1期35-44,共10页Presentday Law Science

基　　金：司法部国家法治与法学理论研究青年课题“网络社群治理的反身法模式研究”(20SFB3003)的阶段性成果。

摘　　要：可解释性决策意味着公共决策需要建立在可被阐明、质疑以及辩护的理由之上,构成了保障民主生活的重要技术方案。近年来崛起的算法决策具有高度的不可解释性,存在理由解释不能、目的解释不能以及逻辑解释不能等问题。算法决策的不可解释性导致了决策结果由“可质疑、可辩护”到“绝对正确”,决策依据从“普遍共识”到“单一意志”,决策机制从“重程序、轻实体”到“重实体、轻程序”,民众身份从“决策主体”到“决策客体”,决策结构从“双向制衡”到“单向控制”的转变,在不同的社会领域再造了算法威权。算法威权的化解有赖于深入的权利保障工作与权力制约工作,通过增进个体在算法决策机制中的对抗力量和对抗机会,强化对算法决策权的约束机制,复原人的社会主体地位。Explainable decision⁃making means that public decision⁃making needs to be based on reasons that can be clarified,questioned,and defended,it constitutes an important technical solution to guarantee democratic life.Algorithmic decision⁃making is unexplainable,there are problems such as the inability to explain the reason,the inability to explain the purpose,and the inability to explain the logic.This has led to a change in decision⁃making results from“questionable and defensible”to“absolutely correct”,decision⁃making basis from“general consensus”to“single will”,decision⁃making mechanism from“emphasis on procedures”to“emphasis on entities”,people􀆳s identity from“decision⁃making subject”to“decision⁃making object”,and decision⁃making structure from“two⁃way checks and balances”to“one⁃way control”,and recreated algorithmic hegemonies in different social fields.The resolution of algorithmic hegemonies depends on in⁃depth rights protection work and power restriction work.We should restore the person􀆳s social subject status by enhancing the individual􀆳s confrontational power and confrontational opportunities in the algorithmic decision⁃making mechanism,and strengthen the restriction mechanism on the decisionmaking power of the algorithm.

关 键 词：民主 算法决策 不可解释 

分 类 号：D90[政治法律—法学理论]