Trump Delays AI Security Executive Order Over 'Blocker' Language

Trump has delayed signing an AI security executive order that would have mandated pre-release government reviews of artificial intelligence models, saying the order's language "could have been a blocker." It's a telling move — the White House wants oversight of AI without actually slowing it down, and those two things don't always fit on the same page.

Background: AI regulation caught between two impulses

Since revoking Biden's AI executive order in January 2025, the Trump administration has sent mixed signals on how it plans to govern this technology. The official line is deregulation — let the market run, keep the U.S. ahead of China. But national security pressure hasn't gone away, and this order was the clearest attempt yet to address it head-on. The tension between those two goals is exactly what apparently killed the draft.

What actually happened

The shelved executive order would have required developers of large-scale AI models to submit them for federal security review before public release — a significant step toward formal government oversight of frontier AI. Trump signed several other orders that same day but held this one back. According to sources close to the administration, the issue wasn't the policy intent but the specific wording, which Trump said could have acted as a brake on industry development. No revised timeline has been announced.

What this really means

This isn't a rejection of AI safety — it's a rejection of language that could be weaponized as a regulatory lever. The administration clearly doesn't want a poorly worded order handing ammunition to critics or creating compliance headaches for companies like OpenAI, Anthropic, and Google DeepMind, all of which have cultivated close ties with the White House. Short-term winner: the tech industry. Short-term loser: anyone hoping for regulatory clarity.

What comes next

The AI sector will continue operating without a federal standard for pre-release security review, at least for now. That means the biggest labs keep setting their own safety benchmarks — which, honestly, is already the norm. It also reduces immediate regulatory pressure on frontier model developers at a moment when the global race for AI dominance is intensifying. On the international stage, the U.S. absence of clear rules makes coordinated global AI governance even harder to pull off.

The real question isn't whether Trump supports AI safety — it's whether any version of this order can be written in a way that doesn't spook the industry the White House is counting on to win the AI race.

Source: TechCrunch