Government Pre-Deployment Testing Now Required for Google, Microsoft, and xAI Models
What Happened
The U.S. Center for AI Standards and Innovation announced agreements with Google DeepMind, Microsoft, and xAI to conduct pre-deployment evaluations of frontier AI models. The program covers capability assessments and targeted research before models reach the public. Anthropic and OpenAI are not part of this initial slate.
My Take
This is the de facto regulatory regime arriving without legislation. Once the government has gated review on three labs, every other lab has to opt in or be branded as the unsafe choice. The smart play for Anthropic and OpenAI is to volunteer for the same review — it's free regulatory moat and it makes Chinese competitors' market access harder. Expect a coordinated announcement within 60 days. The losers are open-weight model providers who can't submit a single canonical version for review.
Read Original Source