| 韩娜,邹初妤.智能认知对抗：理论演化、对抗机制与安全风险[J].国际安全研究,2025,(6):48-69 |
| 智能认知对抗：理论演化、对抗机制与安全风险 |
| Intelligent Cognitive Confrontation: Theoretical Evolution, Confrontation Mechanisms and Security Risks |
| 投稿时间：2025-07-06 修订日期：2025-08-31 |
| DOI：10.14093/j.cnki.cn10-1132/d.2025.06.003 |
| 中文关键词: 智能体 认知对抗 国际博弈 安全风险 |
| 英文关键词: AI agents, cognitive confrontation, international competition, security risks |
| 基金项目: |
|
| 摘要点击次数: 1241 |
| 全文下载次数: 314 |
| 中文摘要: |
| 人工智能催生了新的认知对抗模式，正在重构国际安全秩序并影响认知域安全。从演进维度看，智能认知对抗历经了“智能辅助”的人类主导对抗、“智能赋能决策”的人机协同对抗再到当前“人工智能自主行动”的智能体对抗，已经从单纯的军事作战力量转变成大国竞争的重要战略领域。智能认知对抗在主体、机理和规则方面，呈现由人类主导向算法决策迁移、由线性传播向认知生态系统演化、由信息控制向认知主权侵蚀转变的特征。随着智能体在认知对抗国际实践中的广泛应用，智能体驱动的认知对抗逐渐形成“批判—探索—决策—行动”的循环机制，主要表现为：人工智能通过信息环境动态评估，完成认知叙事网络构建，进而优化竞争战略决策，开展动态认知干预并形成全域智能化的认知对抗工作流。智能体驱动的认知对抗模式正不断引发政治安全的操控风险、社会安全的极化风险、技术安全的失控风险以及文化安全的侵蚀风险。面对上述多维度安全威胁，国家应以系统性视角统筹战略布局，应对安全风险。 |
| 英文摘要: |
| Amid accelerating technological progress and profound international shifts, emerging intelligent technologies—exemplified by AI Agents—have triggered novel forms of cognitive confrontation, thereby redefining the international security order and reshaping threat dynamics within the cognitive domain. From an evolutionary perspective, intelligent cognitive confrontation has escalated from human-dominated “intelligence-assisted” operations, to human-machine collaborative “intelligence-empowered decision-making,” and is now culminating in “AI autonomous action” driven by AI agents. It has transformed from a tactical military operational tool into a core strategic domain in major-country competition. In terms of actors, mechanisms, and rules, this confrontation exhibits distinct characteristics: It marks a shift from human-centric control to algorithmic decision-making, an evolution from linear dissemination to a dynamic cognitive ecosystem, and a transition from information regulation to the erosion of cognitive sovereignty. As AI agents are increasingly integrated into international practices of cognitive confrontation, a cyclic mechanism of Critique-Exploration-Decision-Action has crystallized. Through the dynamic assessment of the information environment, AI facilitates the construction of cognitive narrative networks, which help optimize competitive strategic decisions, enable dynamic cognitive interventions, and finally forge a fully intelligent cognitive confrontation workflow. The AI agent-driven mode of cognitive confrontation continues to generate multiple risks, including the manipulation of political stability, the intensification of social polarization, the loss of control over technological systems, and the erosion of cultural resilience. In response to these multidimensional security threats, China should adopt a holistic perspective to coordinate strategic planning, legal frameworks, and technological innovation, ultimately constructing a comprehensive, dynamic, and sustainable cognitive security governance system to effectively address risks in the cognitive domain. |
|
查看全文
查看/发表评论 下载PDF阅读器 |
| 关闭 |
|
|
|