五角大楼将 Grok 和 ChatGPT 用于军事用途
Pentagon Launches Grok And ChatGPT For Military Use

原始链接: https://www.zerohedge.com/ai/pentagon-launches-grok-and-chatgpt-military-use

美国战争部通过整合两款新的生成式人工智能工具:Starshield AI 的“Grok for Government”和 OpenAI 的“ChatGPT Mil”,扩展了其 GenAI.mil 平台。这两个平台均已获得受控非机密信息(5 级影响)认证,旨在为超过 300 万名人员提升军事生产力、后勤保障及行政效率。 这些新工具的加入反映了五角大楼构建“多模型生态系统”的战略,旨在减少对单一供应商的依赖,并通过安全、专业的 AI 功能加速任务执行。据该部门报告,自九个月前 GenAI.mil 推出以来,已有超过 170 万名人员接入该平台。 此举正值各界对“影子 AI”(即未经授权使用外部 AI 工具)的担忧日益加剧之际。安全机构警告称,此类使用可能导致严重的漏洞和数据泄露风险。五角大楼通过提供获得认可的安全环境,旨在规范联合部队内部的 AI 使用,让人员在保持必要安全保障的同时,能够实现日常任务自动化,从而专注于关键作战行动。

相关文章

原文

Authored by Timothy Frudd via The Epoch Times,

The Department of War announced the launch of two new artificial intelligence (AI) platforms for military use on Aug. 31, expanding the department's platform of AI assistants for military personnel.

Starshield AI's Grok for Government and OpenAI's ChatGPT Mil were both added to the War Department's GenAI.mil generative AI platform on Monday. The Aug. 31 launch came months after the Pentagon announced a partnership with OpenAI to deploy AI models on the Pentagon's classified networks and a partnership with Elon Musk's xAI service to assist personnel with controlled unclassified information.

Announcing the launch of Starshield AI's Grok for Government, the War Department said the AI tool would enable the military to execute missions faster and with more precision in multiple operational contexts. Examples of such contexts include supply chain management for logisticians and market research analysis for acquisition professionals.

The Pentagon said Starshield AI's Grok for Government would provide military personnel with "immediate productivity gains, stronger knowledge continuity, and more secure and efficient collaboration." Capabilities department personnel will have access to include adaptive reasoning modes, customizable workspaces, deep-thinking inference, persistent projects, and reusable "playbooks."

Starshield AI's Grok for Government was accredited for controlled unclassified information at impact level five, a designation given to unclassified information that still requires security safeguards. It was also engineered for "secure, consistent enterprise use," according to the War Department.

With the addition of Starshield AI's Grok for Government to the department's GenAI.mil platform, the Pentagon said military members would have access to another "top-tier generative AI tool." The Pentagon also said the addition of another AI tool would promote a "vibrant" AI ecosystem for the United States and would eliminate its dependence on a single AI provider.

The War Department also announced Monday that it had launched OpenAI's ChatGPT Mil as part of its GenAI.mil platform. Like Starshield AI's Grok for Government, OpenAI's ChatGPT was accredited for controlled unclassified information at impact level five.

"ChatGPT Mil brings a familiar commercial experience into the Department's secure environment, tailored to warfighter needs," the Pentagon said. "The core experience centers on chat, files, projects, and custom GPTs, with additional features sequenced over time."

The department said ChatGPT Mil will support document-heavy unclassified work, including planning, logistics, administration, and policy. Built to support more than 3 million personnel, the AI platform will increase the speed of routine tasks and allow personnel to concentrate on "more critical projects across the Joint Force," the Pentagon said.

"Integrating ChatGPT Mil into GenAI.mil alongside existing frontier AI capabilities establishes a robust, multi-model ecosystem for the warfighter," it added.

The Epoch Times reached out to Starshield AI and OpenAI for comment but did not receive a response before publication time.

The War Department confirmed Monday that more than 1.7 million of its more than 3 million personnel have been onboarded for the department's generative AI platform since GenAI.mil was launched nine months ago.

The launch of the two AI tools for use by War Department personnel comes after the Defense Counterintelligence and Security Agency warned in June that unauthorized "shadow AI" tools could cause data leaks and lead to other security risks.

"Shadow AI encompasses two distinct threat vectors: the intentional use of external commercial or private [large language models], and the activation of embedded AI features within existing government and sensitive networks that have not yet been fully evaluated for security risks," the agency wrote in an assessment.

"When bypassing traditional security controls, both vectors create a massive, unmonitored attack surface where new risks outpace current technical safeguards and governance."

联系我们 contact @ memedata.com