U.S. government to pre-screen latest AI models; Google DeepMind, xAI, MS

The U.S. Commerce Department’s CAISI has struck new agreements with Google DeepMind, Microsoft, and xAI to let the government evaluate frontier AI models before public release.

The setup gives officials access to top-tier models for research and testing, including versions with safety protections reduced when needed to assess risk.

Government evaluators will coordinate with agency experts and provide feedback on national security capabilities and potential hazards.

The move expands pre-release oversight of leading AI systems and strengthens monitoring of both AI safety and security concerns.