Trump Signs New Executive Order to Review AI Models Before Release
Donald Trump has signed a new executive order that creates a voluntary framework for leading AI companies to share advanced model information with the U.S. government before public release.
The goal is to promote safer innovation while strengthening national cybersecurity for important infrastructure. The move marks another attempt to balance rapid AI development with security risks tied to increasingly powerful models.
A Voluntary Framework for AI Companies
The order states that the U.S. AI industry has grown partly because the country avoided overly strict regulation.
However, it also recognizes that new AI capabilities can create security risks. Because of this, government agencies have been directed to create a framework for assessing the advanced cyber abilities of AI models before they are released to the public.
AI companies can still decide for themselves whether they want to share their models or related information with the government. Companies that choose to participate may also receive confidentiality protections.
Review Window Before Public Launch
One important point is the timeline for sharing information.
The current framework allows companies to submit models for review up to 30 days before public release. Major companies such as Google, Microsoft, and xAI have already agreed to cooperate with model review through the relevant government body.
Meanwhile, companies such as OpenAI and Anthropic were already part of earlier safety efforts during the Joe Biden administration.
Not a Full Licensing System
The Trump administration’s approach is described as more flexible than strict regulation.
The order emphasizes that it should not be treated as mandatory licensing or a complete preclearance system. Instead, it shows that the government is becoming more willing to take part in reviewing powerful AI systems after previously taking a more hands-off approach.
One factor behind this change may be the impact of advanced models such as Mythos from Anthropic, which reportedly found many serious security vulnerabilities in major operating systems and web browsers.
Groups Welcome the Move but Want Stronger Rules
The move has received support from several groups that monitor AI risks.
Brad Carson, president of Americans for Responsible Innovation, praised the White House for taking AI vulnerabilities seriously.
Brendan Steinhauser from Alliance for Secure AI also welcomed the move, saying the government is beginning to pay attention to the risks of these powerful models. However, he also called on Congress to create mandatory safeguards in the future.
THIS IS our take
This executive order shows that AI safety is becoming harder for governments to ignore. A voluntary review system may help create trust between AI companies and public agencies without immediately slowing innovation through heavy regulation. Still, as AI models become more powerful, the debate will likely move toward whether voluntary cooperation is enough, or whether stronger legal safeguards will eventually become necessary.
 Origin: The Verge
