• Pentagon ditches Anthropic for Elon Musk’s Grok in massive classified AI shakeup A policy standoff over how military AI can be used is forcing the Pentagon to reconsider its reliance on a single provider for classified operations. • The Pentagon is preparing to expand the artificial intelligence systems running inside its most sensitive classified networks. • Elon Musk’s company xAI has signed an agreement that would allow its Grok model to operate within those secure environments, Axios reported. • The move comes as the Defense Department’s relationship with Anthropic grows increasingly strained. • Until now, Anthropic’s Claude has been the only AI model cleared for use in classified systems tied to intelligence analysis, advanced weapons development, and battlefield operations. • That exclusive access gave Anthropic a central role in some of the military’s most high-stakes workflows.

Article Summaries:

  • The Pentagon is shifting its sole classified‑AI provider from Anthropic’s Claude to Elon Musk’s xAI Grok amid a policy dispute over usage limits. Defense officials want broader access to AI for surveillance and autonomous‑weapon applications, while Anthropic has maintained guardrails that the Pentagon views as restrictive. Grok has agreed to the Department’s standard for lawful use, enabling it to enter the most sensitive networks, though officials question whether it matches Claude’s maturity and battlefield integration. The move follows last year’s approval of Grok, ChatGPT, Gemini, and Claude for government use, with only Claude previously cleared for classified systems.

Sources: