Trump delays AI security executive order, cites concern over language requiring pre-release model review
The president postponed signing an order that would have mandated government evaluation of AI models before their public release, citing competitive concerns and dissatisfaction with proposed requirements for advance model sharing with federal agencies.
1 source · cross-referenced
- Trump delayed an executive order that would require pre-release government security reviews of AI models, stating he did not want language that could impede U.S. competitive leadership in AI.
- The order would have tasked the Office of the National Cyber Director with developing evaluation processes for AI models before release.
- A key point of contention was a proposed requirement for AI companies to submit advanced models to the government 14 to 90 days before public launch.
- Trump told the White House press pool the order's language contained aspects he disliked and could have acted as a 'blocker' on AI development.
President Trump postponed signing an executive order intended to establish government review processes for AI models prior to public release. In remarks to the White House press pool, Trump expressed dissatisfaction with the order's language, stating he did not want measures that would impede U.S. competitiveness against other nations in AI development.
The executive order would have directed the Office of the National Cyber Director and related federal agencies to create and implement a framework for evaluating AI systems' security properties before their commercial or public launch. The timing of the planned order came following high-profile releases of advanced AI systems from Anthropic and OpenAI with heightened security capabilities.
A central provision in the draft order required AI developers to provide pre-release access to advanced models to the government within a 14-to-90-day window before their public announcement. Trump indicated this requirement concerned him as potentially imposing compliance burdens that could hinder innovation velocity. Separately, reporting suggested the delay also reflected scheduling constraints, as key technology sector leaders could not be accommodated for a signing ceremony on the initially planned timeline.
- May 16, 2026 · TechCrunch — AI
ArXiv implements one-year author ban for unchecked AI-generated content
Trust70 - May 2, 2026 · The Verge — AI
Musk testifies xAI used model distillation with OpenAI technology in courtroom proceedings
Trust69 - Apr 28, 2026 · The Verge
Elon Musk testifies in trial against OpenAI leadership over company structure and mission
Trust70