The discretion to share information with the government ultimately rests with AI companies.
President Donald Trump signed an executive order on Tuesday, establishing a "voluntary framework" designed to encourage AI companies to share their advanced models with the federal government prior to public release. This initiative aims "to promote secure innovation and strengthen the cybersecurity of critical infrastructure."
The executive order acknowledges the U.S. AI industry's success, attributing it partly to a rejection of "overly burdensome regulation" that could stifle innovation. However, it simultaneously recognizes the inherent security risks associated with new AI capabilities. Consequently, the order mandates several federal agencies to develop a framework for "assess[ing] the advanced cyber capabilities of AI models" before their public deployment. While companies retain the option to voluntarily share their models with the government pre-release, they would be afforded specific confidentiality protections should they opt in. Furthermore, the directive compels the federal government to bolster its cyber defenses against AI threats, particularly concerning critical infrastructure.
This executive order follows a last-minute postponement by Trump of an earlier, similar directive, which he reportedly feared might "get in the way" of the nation's competitive stance against China. According to The New York Times, the prior version permitted AI companies to voluntarily share their models between 14 and 90 days before public release, whereas the current iteration requests sharing up to 30 days beforehand. Notably, Google, Microsoft, and xAI had already committed last month to allow pre-release scrutiny by the Commerce Department’s Center for AI Standards and Innovation (CAISI).
Under President Joe Biden, OpenAI and Anthropic had already consented to share their models with CAISI in 2024, aligning with Biden's broader initiative for AI safety guardrails. In contrast, the Trump administration, particularly under former White House AI czar David Sacks, had previously minimized safety concerns and adopted a less interventionist stance. The order signed on Tuesday explicitly states that it is not to be interpreted as mandatory licensing or preclearance. Nevertheless, it signifies a notable shift, indicating some readiness by the Trump administration to implement oversight of AI companies.
"The White House is officially Mythos-pilled"
A contributing factor to this policy shift might be Anthropic’s limited release in April of its potent Mythos model. The company reported that Mythos identified "thousands of high-severity vulnerabilities, including some in every major operating system and web browser." Furthermore, the Mythos rollout appears to have created an opportunity to de-escalate tensions between Anthropic and the administration, which had previously been strained by a legal dispute involving the Pentagon concerning the company's AI applications for autonomous lethal weapons and mass surveillance.
The recently signed order has, thus far, received commendation even from organizations that have historically opposed restrictions on state AI legislation. Brad Carson, President of Americans for Responsible Innovation, stated, "The White House is officially Mythos-pilled," further noting that the order demonstrates the Trump administration's commitment to addressing AI vulnerabilities. Similarly, Brendan Steinhauser, CEO of the Alliance for Secure AI, expressed that his group is "pleased to see that the Trump administration is taking the risks of these models seriously." Both Carson and Steinhauser subsequently called upon Congress to enact mandatory protections into law.
The Editorial Staff at AIChief is a team of professional content writers with extensive experience in AI and marketing. Founded in 2025, AIChief has quickly grown into the largest free AI resource hub in the industry.