重庆科技大学学报（社会科学版）Issue(2)：74-84,11.DOI:10.19406/j.issn.2097-4523.2026.02.006
生成式人工智能赋能预测性侦查的正当性证成与限度廓清
Generative AI Empowers the Justification and Limitation of Predictive Investigation
摘要
Abstract
In the context of the frequent occurrence of new economic crimes and cybercrimes,traditional investigation modes have exposed many difficulties,such as complicated investigation procedures and high costs,leading to their weakness in dealing with the crimes above.Based on the interpretation of norms related to the application of generative AI and the empirical reference of extraterritorial predictive policing practice,it is feasible and legitimate for generative AI to be embedded with predictive investigation.However,the practice of generative AI empowering predictive investigation already has difficulties such as weakening of due process,intensification of power alienation,and loss of control of technological risks,which are not unrelated to ambiguous legislative expression and anomie of investigative power.Based on this,it is necessary to strengthen the procedural mechanism of generative AI empowering predictive investigation while refining specific applicable rules around the investigation subjects,investigation objects and initiation standards,and shaping a technical guarantee mechanism with algorithm filing and algorithm anonymization mechanism as the core.关键词
生成式人工智能/预测性侦查/预测性警务/算法备案/算法匿名化Key words
Generative AI/predictive investigation/predictive policing/algorithm filing/algorithm anonymization分类
社会科学引用本文复制引用
曾德梅..生成式人工智能赋能预测性侦查的正当性证成与限度廓清[J].重庆科技大学学报（社会科学版）,2026,(2):74-84,11.基金项目
广东省普通高校青年创新人才类项目"心理测试技术在监狱工作中的应用"(2019GWQNCX062) （2019GWQNCX062）
广东省普通高校特色创新类项目"数字化转型背景下警务特色虚拟仿真教学资源开发与应用"(2023WTSCX223). （2023WTSCX223）
