Donald Trump postpones AI security executive order

President Donald Trump announced he delayed signing an executive order aimed at addressing cybersecurity risks from advanced artificial‑intelligence models. He said he “didn’t like certain aspects” of the draft and wanted to avoid anything that could hinder the United States’ technological lead over China.

The proposed order would have expanded existing voluntary information‑sharing programs to include AI companies, allowed government testing of frontier AI systems for vulnerabilities, and encouraged major tech firms—such as Google, Microsoft, xAI, OpenAI and Anthropic—to share their models with the Commerce Department’s Center for AI Standards and Innovation. The directive stopped short of imposing mandatory federal approval for cutting‑edge models, instead calling for voluntary testing across federal, state, local networks and critical infrastructure.

Industry leaders have already participated in related voluntary assessments, and the National Security Agency has used Anthropic’s Mythos model to probe network weaknesses. The postponement injects uncertainty into the administration’s broader AI policy push, which seeks to balance rapid innovation with national‑security safeguards.