AI News AI资讯 20h ago Updated 15h ago 更新于 15小时前 49

US military adds ChatGPT and Grok to AI platform GenAI.mil 美军将ChatGPT和Grok加入AI平台GenAI.mil

The Pentagon is adding OpenAI's ChatGPT Mil and xAI's Grok for Government to its GenAI.mil platform, expanding beyond the previously sole offering of Google Gemini The platform now serves over 1.7 million users out of more than 3 million DoD employees and military personnel Both models operate in a dedicated secure environment isolated from commercial versions, with no user data collection ChatGPT Mil is positioned for administrative tasks, logistics, and planning, while Grok targets procurement 五角大楼GenAI.mil平台新增ChatGPT Mil和Grok for Government两个AI模型,此前仅集成Google Gemini 平台用户已达170万,覆盖超300万军事人员和员工,体现大规模政府AI部署 两个工具均在隔离安全环境中运行,与商业版本物理隔离,不收集用户数据 Anthropic Claude被排除在外,因拒绝授予无条件使用权限,供应链风险分类被法院裁定非法

75
Hot 热度
65
Quality 质量
70
Impact 影响力

Analysis 深度分析

TL;DR

  • The Pentagon is adding OpenAI's ChatGPT Mil and xAI's Grok for Government to its GenAI.mil platform, expanding beyond the previously sole offering of Google Gemini
  • The platform now serves over 1.7 million users out of more than 3 million DoD employees and military personnel
  • Both models operate in a dedicated secure environment isolated from commercial versions, with no user data collection
  • ChatGPT Mil is positioned for administrative tasks, logistics, and planning, while Grok targets procurement analysis and supply chain management
  • Anthropic's Claude remains excluded after being classified as a supply chain risk by the Trump administration, though a court later ruled that classification unlawful

Why It Matters

The expansion of GenAI.mil signals the DoD's accelerating adoption of commercial AI models for defense operations, reflecting a broader institutional shift toward integrating third-party LLMs into government workflows. The exclusion of Claude despite a court ruling highlights the ongoing tension between AI vendors' data governance policies and government security demands.

Technical Details

  • GenAI.mil now hosts three foundational models: Google Gemini (available since December), OpenAI's ChatGPT Mil, and xAI's Grok for Government, each deployed in isolated secure environments
  • Both ChatGPT Mil and Grok for Government are walled off from their commercial counterparts to prevent sensitive government data exposure, with zero user data collection policies
  • ChatGPT Mil is optimized for administrative tasks, logistics coordination, and planning workflows
  • Grok for Government is positioned for procurement analysis and supply chain management, with military-specific use cases emphasized
  • Anthropic's Claude was previously excluded due to the company's refusal to grant unrestricted Pentagon access, leading to its supply chain risk classification

Industry Insight

The DoD's multi-vendor strategy on GenAI.mil demonstrates that government agencies are moving away from single-provider dependencies, creating a more competitive landscape for AI vendors seeking defense contracts. The Claude exclusion episode underscores that data sovereignty and vendor cooperation terms will remain critical gatekeepers for AI access to sensitive government sectors, regardless of judicial outcomes.

TL;DR

  • 五角大楼GenAI.mil平台新增ChatGPT Mil和Grok for Government两个AI模型,此前仅集成Google Gemini
  • 平台用户已达170万,覆盖超300万军事人员和员工,体现大规模政府AI部署
  • 两个工具均在隔离安全环境中运行,与商业版本物理隔离,不收集用户数据
  • Anthropic Claude被排除在外,因拒绝授予无条件使用权限,供应链风险分类被法院裁定非法

为什么值得看

五角大楼大规模部署商业AI模型,标志着政府机构对AI技术的采纳进入新阶段。这一举措反映了军事领域对AI工具的实际需求,同时也揭示了AI安全与国家安全之间的复杂关系。

技术解析

  • GenAI.mil平台此前仅集成Google Gemini,现扩展至OpenAI和xAI的政府定制版本,形成多模型架构
  • ChatGPT Mil专注于行政、后勤和规划任务,Grok则针对采购分析和供应链管理,体现不同模型在军事场景中的专业化分工
  • 两个工具运行在完全隔离的安全环境中,与商业版本物理分离,确保敏感政府数据不会泄露,且平台不收集任何用户数据
  • Anthropic Claude被排除在平台之外,原因是公司拒绝授予五角大楼unrestricted use权限,特朗普政府曾将其列为供应链风险,但法院在8月底裁定该分类非法

行业启示

  • 政府机构正在加速采用商业AI模型,但安全隔离和数据保护成为关键考量因素,供应商需满足严格的政府合规要求
  • AI供应商需要在商业利益和国家安全需求之间找到平衡,Claude的排除案例显示了这种张力,可能影响其他政府合作
  • 多模型策略正在成为政府AI平台的标准做法,不同模型针对不同任务场景进行专业化部署,避免单一供应商锁定

Disclaimer: The above content is generated by AI and is for reference only. 免责声明:以上内容由 AI 生成,仅供参考。

GPT GPT LLM 大模型 Product Launch 产品发布 Security 安全 Policy 政策