政府部门人工智能大模型应用安全研究  

Research on the Safety and Cyber-security of Government Artificial Intelligence Applications

作　　者：张瑞宏 苗伟彬 高志国 李钊 陈奋 曹辉 ZHANG Ruihong;MIAO Weibin;GAO Zhiguo;LI Zhao;CHEN Fen;CAO Hui(Jinan Customs,Jinan Shandong 250001,China;Yaxin Security Technology Co.,Ltd,Beijing 100176,China;Tsinghua University,Beijing 100084,China;Wuhan University,Wuhan Hubei 430072,China)

机构地区：[1]济南海关,山东济南250001 [2]亚信安全科技股份有限公司,北京100176 [3]清华大学,北京100084 [4]武汉大学,湖北武汉430072

出　　处：《口岸非传统安全学刊》2025年第2期72-79,共8页

基　　金：济南海关科研项目(2023JK014)。

摘　　要：本文介绍了人工智能大模型发展及应用情况,列举了近年来与人工智能大模型相关的政府网络安全事件,分析人工智能大模型在全生命周期安全、底座安全和供应链安全方面存在的风险点,并给出人工智能大模型安全应用建议,通过强化风险评估与威胁分析,构建大模型安全技术体系,实施综合治理,推动标准化与开展多方合作等方式,实现政府部门人工智能大模型的安全应用。This article introduces the development and application of Artificial Intelligence large model,lists recent government cybersecurity incidents related to Artificial Intelligence large model,analyzes the risk points in the full life cycle of Artificial Intelligence large model,bottom-up security,and supply chain security,and provides recommendations for the safe application of Artificial Intelligence large model.By strengthening risk assessment and threat analysis,building a secure technology system for Artificial Intelligence large model,implementing comprehensive governance,promoting standardization,and conducting multi-party cooperation,government departments can achieve safe application of Artificial Intelligence large model.

关 键 词：人工智能大模型 政府部门 网络安全数据安全 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]