Big Tech Hands US Government Early Access to Cutting-Edge AI Models for National Security
xAI, Google, and Microsoft have struck a deal to give the US government pre-release access to their most advanced frontier AI systems. The goal: thorough risk assessments on national security threats before these models hit the public.
The program runs through CAISI, the Center for AI Standards and Innovation at the Commerce Department’s NIST.
It already completed over 40 evaluations covering cybersecurity, biosecurity, and other high-stakes dangers.
This expands voluntary partnerships that included OpenAI and Anthropic, putting America’s top AI builders in direct coordination with Washington to stay ahead of adversaries.