White House prepares executive order for voluntary government review of AI models before public release

The White House plans to issue an executive order as soon as Thursday requiring major AI companies to submit advanced models for government review before release. The review period is disputed-drafts range from 90 days to 14 or fewer.

Categorized in: AI News Government
Published on: May 21, 2026
White House prepares executive order for voluntary government review of AI models before public release

White House Readies AI Model Review Order for Government

The White House plans to issue an executive order as soon as Thursday that would push major AI companies to voluntarily submit advanced models for government review before public release, according to sources briefed on the effort.

The order includes a voluntary agreement where AI firms would share models with federal agencies ahead of launch. The review period remains contested: one draft proposes 90 days, while companies like OpenAI and Anthropic have pushed for 14 days or less.

What the Order Would Cover

The executive order splits into two sections. One addresses cybersecurity vulnerabilities in unreleased models. The other, titled "covered frontier models," defines which AI systems qualify for early government access and sets the terms.

The cybersecurity section calls for a voluntary "clearinghouse" run by the Treasury Department and other agencies to identify and fix security flaws before models reach the public. It also directs increased hiring at the US Tech Force, which modernizes federal computer systems.

Why the Government Wants Early Access

Advanced AI models can amplify cyberattacks at scale, security experts have said. Early government review could help agencies defend against threats before the models become widely available.

The timing reflects recent developments in the field. Anthropic unveiled its Mythos model, which the company says can identify cybersecurity vulnerabilities faster than previous systems. Anthropic has not released it publicly and instead granted access through Project Glasswing, a controlled consortium. OpenAI similarly provides early access to its latest models for government and business cybersecurity work.

Industry Engagement

Major AI companies have been negotiating with White House officials on the order's terms. OpenAI and Anthropic are among those involved in discussions about the review framework and timeline.

The National Institute of Standards and Technology announced earlier this month that major tech firms would share unreleased models with the government for national security evaluation, though that announcement has since been removed from the Commerce Department's website.

The White House declined to comment on the executive order.


Get Daily AI News

Your membership also unlocks:

700+ AI Courses
700+ Certifications
Personalized AI Learning Plan
6500+ AI Tools (no Ads)
Daily AI News by job industry (no Ads)