AI Briefing
KO

US conducts national security review of AI models

·2026.05.06 06:46

Key point

The US CAISI will pre-review pre-release AI models from Google DeepMind, Microsoft, and xAI against national security standards.

Details

The CAISI, under the US Department of Commerce, has signed agreements with Google DeepMind, Microsoft, and xAI to pre-review pre-release AI models from a national security perspective.

  • Review scope: cybersecurity, biosecurity, chemical weapons risk
  • Purpose: to more accurately measure the capabilities and risks of frontier models and strengthen public safety
  • Status: over 40 evaluations have already been conducted, including of unreleased models.

This agreement extends the existing pattern of cooperation with OpenAI and Anthropic. Microsoft separately signed a similar agreement with the UK's AI Security Institute as well.

national-securityai-policyai-governance
Read originalShare on Threads

This summary was generated automatically by AI. Check the original for the author's claims and context. Copyright belongs to the original author.

Our guide explains how the AI works. Report summary errors, attribution issues, or removal requests via Contact.

Related

Hacker NewsUS Appeals Court Upholds Pentagon Blacklisting of Anthropic as Supply Chain RiskSimon WillisonOpen Letters on AI Development ProliferateAnthropic NewsAdditional $20 Million Donation to Public First ActionRedditUS and China Hold First AI Dialogue; US Proposes Incident Notification MechanismGeekNewsAustralian PM Reports OpenAI Agent Hacked Medicare Portal in Believed First Known AI-Driven Government BreachTLDR AIHugging Face CEO details lessons from first public agent cyberattack, advocates for open-source defense