美国特种作战司令部今年春季在与伊朗战争期间经历了一次危险的误判1。一名分析人员使用聊天机器人处理情报时，该工具错误识别了一艘中国船只的货物，生成了声称该船在中东运输核武器项目组件的情报报告1。美军随即制定了拦截计划，军机升空并准备武装人员登船，直至最后时刻才发现该报告完全由AI生成且充满虚假信息1。这一虚假情报险些引发美中之间的军事冲突1。
美军目前正在情报工作、目标选择、后勤等几乎所有方面推进AI应用1。防长Pete Hegseth于1月发布了《人工智能加速战略》，旨在向300万名军事和文职人员提供AI工具1。然而，不同部门使用不同的AI工具，缺乏统一的信息验证标准1。此外，AI在目标瞄准中的应用还存在缺乏人类监督机制以防止平民伤亡的问题1。
The US military experienced a dangerous near-miss involving artificial intelligence this spring when a Special Operations Command analyst used a chatbot to process intelligence, resulting in a fabricated report claiming a Chinese vessel was transporting nuclear weapons components in the Middle East.1 Based on this AI-generated assessment, military officials developed an interception plan, deploying aircraft and readying armed personnel to board the ship before discovering at the last moment that the entire report was false and AI-generated.1 The incident nearly sparked an armed conflict between the United States and China.12
The episode underscores broader vulnerabilities in the Pentagon's rapid AI integration across military operations.1 Defense Secretary Pete Hegseth released an "Artificial Intelligence Acceleration Strategy" in January aimed at providing AI tools to 3 million military and civilian personnel.1 Currently, different military departments deploy different AI systems without unified information verification standards, and AI applications in targeting lack human oversight mechanisms to prevent civilian casualties.1 The military has already advanced AI applications across intelligence work, target selection, logistics, and nearly all other operational domains.1
评论
还没有评论，欢迎留下第一条。