US government will get early access to AI models from Google, Microsoft and xAI

Google, Microsoft, and xAI (Elon Musk's AI company) reached an agreement with the U.S. government to share their new AI models before they launch to the general public.

The goal is for an agency of the U.S. Department of Commerce — specifically the Bureau of Industry and Security — to evaluate the safety of these systems before they become available to everyone. The process seeks to identify potential risks before models reach millions of users.

What does the agreement involve?

Companies committed to providing early access to advanced AI models when they exceed certain capability thresholds. The government agency will be able to run them, test them, and request changes before commercial distribution is permitted.

This agreement represents a significant change in how AI releases are supervised in the United States. Previously, companies decided when and how to launch their models without prior government review. Now, for the most advanced systems, there will be a formal evaluation process.

The companies affected

The agreement includes the three companies currently leading AI model development in the United States: Google (with Gemini), Microsoft (with its partnerships with OpenAI and its own models), and xAI (with Grok). Other companies in the sector will be evaluated for future inclusion.

Perspective

The U.S. government is building a supervision framework for AI that didn't exist before. It's a move that reflects how AI ceased to be just a software product to become a technology with national security implications.