President Donald Trump has signed an executive order directing leading AI developers, including Google, OpenAI, and Anthropic, to voluntarily submit their most capable models for government cybersecurity testing before public release. The order gives agencies up to 30 days to test the models.
The directive marks a shift in Trump's approach to AI regulation, moving from a hands-off stance to active monitoring. While voluntary, the testing could slow product rollouts and impact industry profits if companies adjust operations to address security concerns.
The order also tasks the Treasury Secretary with establishing an AI cybersecurity clearinghouse, developed in collaboration with the AI industry, to coordinate vulnerability scanning and patch distribution.