刑事审判人工智能的权力逻辑  

The Power Logic of Artificial Intelligence in Criminal Trials

作　　者：卫晨曙 WEI Chenshu(Institute of Evidence Law and Forensic Science,China University of Political Science and Law,Beijing 100088,China)

机构地区：[1]中国政法大学证据科学研究院,北京100088

出　　处：《西安交通大学学报（社会科学版）》2021年第3期141-148,共8页Journal of Xi'an Jiaotong University：Social Sciences

基　　金：教育部规划基金项目(19YJA820004)。

摘　　要：刑事智能办案系统是人工智能技术进入刑事司法场域中的主要表现形式。当前,刑事智能办案系统主要集中在证据审查判断、辅助精准量刑、类案推送和案件偏离预警等领域。随着人工智能技术与刑事司法的深度融合,人的“代具性”原理、自动化偏差认知机制、国家权力的背书与支持、技术外包的研发模式以及系统架构的运行机制等因素推动着智能办案系统从“工具”向“权力”的演化。同时,刑事审判中人工智能技术的权力化可能会引发规训风险、排斥风险与误判风险。为了有效规避刑事审判场域中人工智能技术引发的正当性风险,应当树立权力规制理念,从智能办案系统的适用机制、参与机制以及研发机制三方面进行规制。Criminal intelligence case handling system is the main manifestation of artificial intelligence technology in the criminal justice field.At present,the criminal intelligence case handling system is mainly concentrated in the fields of evidence review and judgment,auxiliary accurate sentencing,delivery of category cases and warning of case deviation.As the deep integration of artificial intelligence technology and criminal justice,the“prothèse”of technological philosophy,automation bias recognition mechanism,endorsement and support of state power,the model of technology outsourcing and system architecture and so on are driving the transformation of the intelligent case handling system from“tools”to“power”.At the same time,the power of artificial intelligence technology in criminal trials may cause disciplinary risks,rejection risks,and misjudgment risks.In order to effectively avoid the legitimate risks caused by artificial intelligence technology in the criminal trial,the idea of power regulation should be established,and the three aspects of application mechanism of criminal intelligent case handling system,participation mechanism and research and development mechanism should be regulated.

关 键 词：刑事审判 人工智能 技术权力 智能办案 正当程序 司法裁判权 

分 类 号：TP18[自动化与计算机技术—控制理论与控制工程] D925.2[自动化与计算机技术—控制科学与工程]