en.Wedoany.com Reported - The White House Office of the National Cyber Director held a briefing for artificial intelligence companies on May 19, outlining a planned executive order on AI. The order would authorize intelligence agencies and other government bodies to review advanced AI models before their public release. Companies attending the briefing included OpenAI, Anthropic, and Reflection AI, and U.S. President Trump could sign the order as early as the 21st. Representatives from cloud service providers, semiconductor companies, cybersecurity firms, and banks were also present.
Under the proposed framework, AI labs would be required to submit new models to the government up to 90 days before public release, allowing intelligence agencies and other government bodies to assess potential national security risks. This framework is built on voluntary principles, establishing a standardized pre-release notification process for frontier model developers.
This proposed executive order marks a significant shift in the Trump administration's AI regulatory policy. In early May, the White House had already begun formulating related plans, discussing the signing of an executive order to establish an AI task force composed of government officials and tech company executives. This task force would study various regulatory pathways, including introducing a government review mechanism before the release of new AI models. Executives from companies like Anthropic have already communicated with the White House regarding these ideas.
The draft executive order contains at least two major sections. The first focuses on cybersecurity, requiring the Pentagon to secure its networks within 30 days, covering critical telecommunications and information systems. The second section focuses on the definition and review process for "controlled frontier models," designating a coalition of national security and civilian agencies to strengthen scrutiny of advanced AI models.
During the policy formulation process, White House National Economic Council Director Hassett publicly stated that the government is studying a possible roadmap for executive orders, hoping that AI will be proven safe before being released into the public domain.
The accelerated push for this executive order is closely tied to the recent release of Anthropic's Mythos model. Mythos's demonstrated vulnerability discovery capabilities in the cybersecurity field have raised significant concerns within government agencies about the potential risks of AI models, prompting the White House to speed up the establishment of a model review mechanism.
The corporate representatives attending the briefing spanned the entire AI technology stack—from AI model developers OpenAI, Anthropic, and Reflection AI, to cloud service providers, semiconductor companies, cybersecurity firms, and banks. This lineup indicates that the potential impact of the executive order will extend from model development to multiple industrial segments, including computing power supply, application deployment, and financial services.
Once the executive order is officially signed, the U.S. federal government will intervene in the safety assessment process before the commercial deployment of frontier AI models, forming a complete framework covering pre-release review, interagency coordination, and voluntary corporate notification. Against the backdrop of a rapidly evolving global AI regulatory landscape, this policy move will provide a reference point for the AI governance pathways of other major economies.
This article is compiled by Wedoany. All AI citations must indicate the source as "Wedoany". If there is any infringement or other issues, please notify us promptly, and we will modify or delete it accordingly. Email: news@wedoany.com
