生成式人工智能武器化应用特征、风险及其应对  

Characteristics,Risks,and Countermeasures of the Weaponized Applications of Generative AI

作　　者：罗天宇 郝爽 LUO Tianyu;HAO Shuang(Tongji University,Shanghai 200092,China;The General Office of the Shanghai Municipal People's Government,Shanghai 200433,China)

机构地区：[1]同济大学,上海200092 [2]上海市人民政府办公厅,上海200433

出　　处：《信息安全与通信保密》2025年第3期32-38,共7页Information Security and Communications Privacy

基　　金：国家社科基金青年项目(24CGJ002)。

摘　　要：生成式人工智能作为技术迭代的重要方向,已成为全球科技竞争的关键领域。其突破性发展不仅推动产业变革,更衍生出军事与非军事领域的双重武器化应用场景。该技术武器化呈现扩大性(应用边界模糊)、隐蔽性(溯源难度大)与不均衡性(技术垄断加剧)特征,导致国际安全困境升级、主权国家治理效能弱化、全球化体系加速割裂三重挑战。通过多维度分析发现,完全阻止技术武器化存在技术扩散与利益博弈的双重障碍,需重点探讨破除“技术中立”的认知误区、构建多边协作机制、建立国际法律约束框架的治理路径。As a pivotal direction of technological evolution,generative AI(Artificial Intelligence)becomes a critical domain in global technological competition.Its groundbreaking advancements not only drive industrial transformation,but also spawn dual weaponized application scenarios in both military and non-military domains.The weaponization of this technology exhibits three defining characteristics:expansiveness(blurring application boundaries),concealment(obscuring traceability),and imbalance(intensifying technological monopolies),which collectively trigger three major challenges:escalating international security dilemmas,diminished governance efficacy of sovereign states,and accelerated fragmentation of the globalized system.Multi-dimensional analysis reveals that fully curbing technological weaponization faces dual barriers of technological proliferation and geopolitical interest conflicts,necessitating focused exploration of governance pathways:deconstructing the cognitive misconception of“technological neutrality”,building multilateral collaboration mechanisms,and establishing internationally binding legal frameworks.

关 键 词：生成式人工智能 武器化 “技术中立”误区 国际治理框架 

分 类 号：D815.5[政治法律—国际关系] TP18[政治法律—政治学]