White House Launches New AI Security Framework
President Donald Trump has issued a new executive order aimed at maintaining United States AI leadership while addressing the security risks posed by increasingly powerful AI systems.
The latest version of the order creates a voluntary review process for advanced AI models, focuses on cybersecurity and critical infrastructure, and avoids heavy regulation.
Trump had refused to sign an AI order on May 21 after raising concerns that it could leave the U.S. vulnerable to losing ground to China in the AI race, the Associated Press reported.
Under the order, AI companies may provide their most advanced models to the federal government before public release, allowing officials to evaluate potential security risks. The review process is limited to 30 days.
The order also calls for the creation of a cybersecurity clearinghouse to share information on AI-enabled risks and vulnerabilities, stating, "Advanced AI capabilities make our Nation stronger, but also introduce new national security considerations."
Another major talking point of the order is free rein for AI companies. The order seeks to loosen the noose on excessive regulation or licensing requirements that the government says could slow innovation and competitiveness. Crucially, participation in the review process is voluntary rather than mandatory.
"We need to strike the right balance between accelerating U.S. AI innovation and ensuring advanced models are tested for risks before broader deployment," said Katharina Sommer, group head of Government Affairs at cybersecurity consultancy NCC Group. "Voluntary guidance alone is not enough. Current risk levels, combined with the acceleration of frontier models, mean relying solely on non-binding advice would leave systemic vulnerabilities unmanaged."
"In short, we should not sacrifice security in the name of speed, but avoid regulatory 'freeze' by using staged or adaptive obligations supported by global best practice frameworks," Sommer said.
The big U.S.-based AI companies, such as Anthropic, OpenAI, and Google, known as frontier labs, were hoping to avoid mandatory government approval before releases.
On the other hand, national security officials were concerned that frontier AI models could identify vulnerabilities, accelerate cyberattacks, or create other security risks.
The executive order provides insight into how the White House is hoping this framework can help the US lead the AI race and manage AI risks simultaneously, without imposing heavy-handed regulation on the industry's biggest players.
The full executive order is available here on the White House site.