International Open Access Journal Platform

人机协同视角下治安处罚裁量辅助系统的技术原理、风险检视与应对策略

Technical Principles, Risk Review and Response Strategies of Public Security Punishment Discretionary Assistance System from the Perspective of Human-Machine Collaboration

Criminal Justice Science & Governance / 2025,6(2): 113-122 / 2025-12-05 16 7

View Download PDF

• Authors: 常克强
• Information:

  中南财经政法大学刑事司法学院，武汉

• Keywords:
  Human-machine collaboration theory; Computational law; Technical principle; Risk review; Response strategy
  人机协同理论; 计算法学; 技术原理; 风险检视; 应对策略
• Abstract: The integration of artificial intelligence technology and public security law enforcement has become an inevitable trend. Through the field research on Wuhan DearSeeWe Intelligent Technology Co., Ltd, this paper finds that current public security law enforcement is evolving from the “substantive decision-making” stage to the “substantive prediction” stage, and the technical application is also transforming from “expertism” to “connectivism”. However, most existing academic research focuses on the legal and procuratorial systems and “expertism”-related fields, and research on the application of “connectivism” technology in public security law enforcement is still blank. Taking this as the entry point, this paper first clarifies the technical principles and construction connotation of the public security punishment discretionary assistance system; second, taking correlation and causality, subjectivity and objectivity, independence and dependence as the core contradiction origins, it summarizes three major risk types recognized in the theoretical circle: procedural risks, substantive risks and algorithmic risks; finally, combined with the new application background of legal large models, it re-examines the above risks, and supported by the “human-machine collaboration” theory, puts forward corresponding risk response paths from three dimensions: application strategy, algorithmic strategy and procedural strategy, aiming to realize the organic combination of technical advantages and the professional advantages of law enforcement personnel and achieve a win-win situation. 人工智能技术和公安执法的结合势不可挡。本文通过对武汉迪赛威智能科技有限公司的实地调研发现，目前公安执法正从“实体决策”向“实体预测”阶段演进，技术应用也从“专家主义”向“连接主义”转型。然而，学界现有研究多聚焦于法检系统和“专家主义”相关领域，针对公安执法中“连接主义”技术应用的研究尚属空白。以此为切入点，本文首先明晰了技术的原理与构建内涵；其次以相关性与因果性、主观性与客观性、独立性与依赖性为核心矛盾原点，归纳出理论界公认的程序风险、实体风险与算法风险三大风险类型；最后结合法律大模型的新型应用背景，对上述风险再次审视，并以“人机协同”理论为支撑，从应用策略、算法策略与程序策略三个维度提出相应风险应对路径，旨在实现技术优势和执法人员专业优势的有机结合，达成共赢局面。
• DOI: https://doi.org/10.35534/cjsg.0602012
• Cite: 常克强．人机协同视角下治安处罚裁量辅助系统的技术原理、风险检视与应对策略［J］．刑事司法科学与治理，2025，6（2）：113-122．