Trump's executive order creates voluntary framework for AI companies to submit models for government safety testing before release

President Trump's executive order directs federal agencies to strengthen cyber defense of government systems within 30 days and tasks the Department of Defense, CISA, and the Treasury Department with setting up a clearinghouse for software vulnerabilities in partnership with the AI industry.

The order establishes a voluntary framework for "covered frontier models" where AI developers can submit models to the government for safety testing before release, explicitly ruling out any mandatory approval process.

Google DeepMind, Microsoft, and xAI have already agreed with the US Center for AI Standards and Innovation (CAISI) to submit new AI models for review ahead of release; OpenAI and Anthropic entered a similar agreement in 2024.

OpenAI published a strategy paper calling for a binding national AI safety framework, mandatory pre-release testing of the most powerful models by CAISI, independent audits, and whistleblower protections.