算法预测性警务的风险辨析及应对  

Risk Identification and Management of Algorithmic Predictive Policing

作　　者：冯伊蜓 FENG Yiting(School of Law,Lanzhou University,Lanzhou Gansu 730000)

机构地区：[1]兰州大学法学院,甘肃兰州730000

出　　处：《长治学院学报》2024年第1期43-50,共8页Journal of Changzhi University

摘　　要：随着算法时代的到来,在总结吸收问题导向警务和热点警务模式特性的基础上,通过对“不安全因素”的边际判断以及算法决策机制,算法预测性警务实现了对警情的预报预测预警。但算法预测性警务也存在风险,包括计算机学习的样本缺陷、技术上的利维坦化以及算法黑箱在内的技术风险,还包括算法在嵌入警察权的过程中可能带来的警察权扩张、“全景敞视主义”的自由困境、破坏警民关系等制度风险。为此,除了扩大数据池以及加强计算机学习模型的优化等技术升级保障外,更要做好个人信息保护与国家监管之间、算法预测的工具理性与伦理道德之间、算法技术垄断与公民参与之间三方面的平衡,以更好地应对算法预测性警务在未来可能面对的风险。With the advent of the algorithms era,Algorithmic Predictive Policing realises the predictions of police circumstance,based on summarizing and assimilating the characteristics of the Problem-oriented Policing and Hot Spots Policing models,through marginal judgments of“insecurity factors”and algorithmic decision-making mechanisms.However,there are also risks associated with Algorithmic Predictive Policing,including technical risks due to sample defects in computer learning,technological Leviathanization,and algorithmic black-box,as well as institutional risks such as the expansion of police power,the dilemma of freedom of“panoramic openness”,and the undermining of police-citizen relations that algorithms may bring about when embedded in the process of police power.In this regard,in addition to expanding the data pool and strengthening the optimization of computer learning models and other technological upgrades,it is necessary to strike a good balance between the protection of personal information and state regulation,the instrumental rationality of algorithmic prediction and ethics,algorithmic technological monopoly and civic participation,in order to better cope with the risks that algorithmic predictive policing may face in the future.

关 键 词：预测性警务 预测性警务底层逻辑 算法技术风险 风险规制 

分 类 号：D631[政治法律—政治学] TP18[政治法律—中外政治制度]