首页|期刊导航|情报杂志|五间并用:基于AI大模型的认知域安全风险预警体系构建

五间并用:基于AI大模型的认知域安全风险预警体系构建OACHSSCD

Five Rooms in Use:Construction of a Cognitive Domain Safety Risk Early Warning System Based on Large AI Models

中文摘要英文摘要

[目的]随着数智技术的迭代发展,国家认知域安全面临严峻复杂挑战,深入剖析当前认知域安全风险的演化路径及典型特征,能够不断提升我国认知安全能力和韧性治理水平.[方法]通过引入我国古代"五间并用"的情报思想,借助AI大模型的性能优势和算力支持,在保留OODA循环理论的结构基础之上,构建以"情报察觉牵引—态势感知定位—韧性决策驱动—系统恢复成长"为核心的认知域安全风险预警模型.[结果/结论]研究发现,"五间"思想与AI大模型的深度融合与现代转译,能够有效克服传统预警体系信息要点遗漏、风险识别不足以及事实幻觉等问题,实现了对认知域安全风险的精准识别、动态评估、科学防控和全面应对,完成了从"情报优势"向"认知优势"的转变,为服务国家安全体系与能力现代化建设提供理论与实践层面的支撑.

[Purpose]With the iterative development of digital-intelligent technology,national cognitive domain security faces severe and complex challenges.In-depth analysis of the evolutionary pathways and typical characteristics of current cognitive domain security risks can further enhance the country's cognitive security capabilities and resilience governance level.[Method]By introducing China's ancient intelligence concept of"five-way collaboration",leveraging the performance advantages and computing power support of large AI mod-els,and retaining the structural foundation of the OODA cycle theory,a cognitive domain security risk early warning model centered a-round"intelligence awareness traction—situational awareness positioning—resilience decision-making drive—system recovery growth"is constructed.[Result/Conclusion]The study finds that the deep integration and modern translation of the"five-way"concept with AI large models can effectively overcome issues such as information key points oversight,insufficient risk identification,and factual hallucina-tions in traditional early warning systems.It achieves precise identification,dynamic assessment,scientific prevention,and comprehensive response to cognitive domain security risks,completing the transformation from"intelligence advantage"to"cognitive advantage",and providing theoretical and practical support for the modernization of the national security system and capabilities.

吉博阳;李勇男

中国人民公安大学国家安全学院 北京 100038中国人民公安大学国家安全学院 北京 100038||智慧警务与国家安全风险治理重点实验室 泸州 646000

社会科学

认知域安全五间并用AI大模型风险预警国家安全治理情报智能决策

cognitive domain securityfive-way collaborationlarge AI modelsrisk early warningnational security governanceintelli-gence-driven decision making

《情报杂志》 2026 (4)

49-57,83,10

公安部重点项目(编号:2023JSZ04)中国人民公安大学基本科研业务费项目"大数据背景下国家安全情报风险态势感知模型研究"(编号:2024JKF01)智慧警务与国家安全风险治理重点实验室开放课题"城市公共安全风险态势感知与防范应对研究"(编号:ZHKFYB2502)研究成果.

10.3969/j.issn.1002-1965.2026.04.007

评论