Pre-Deployment AI Evaluation Moves From China’s Model To Washington

The U.S. Commerce Department’s CAISI will evaluate frontier AI systems from Google DeepMind, Microsoft, and xAI before public release.

The move marks a shift toward pre-deployment testing, bringing U.S. AI oversight closer to China’s pre-release review model.

Officials frame the policy as a national security and cybersecurity safeguard for powerful AI models.

Supporters say it could improve safety, while critics worry about regulatory overreach and slower innovation.