Background: An Executive Order Takes Shape
<cite index="19-2">In its June 2, 2026 Executive Order (EO) 14409, "Promoting Advanced Artificial Intelligence Innovation and Security," the White House directed federal agencies to form an artificial intelligence (AI) cybersecurity clearinghouse in voluntary collaboration with the AI industry and operators of critical infrastructure.</cite> <cite index="16-4,16-5">The order asks companies, on a voluntary basis, to participate in a benchmarking process to assess a model's "advanced cyber capabilities" and determine whether it should be considered a "covered frontier model," and requests access to those models up to 30 days before the companies plan to release them more broadly, enabling the government to help select the "trusted partners" that will receive early access.</cite> Notably, <cite index="16-6">the order itself states that "nothing in this section shall be construed to authorize the creation of a mandatory governmental licensing, preclearance, or permitting requirement for the development, publication, release, or distribution of new AI models, including frontier models."</cite>
Gold Eagle: Structure and Operations
<cite index="19-3,19-4">The anticipated AI cybersecurity clearinghouse took shape on July 14, 2026, with the White House's announcement of the launch of "Gold Eagle." According to the White House's announcement, the clearinghouse is designed to enable cybersecurity vulnerability coordination across the federal government and industry by leveraging frontier AI to intake, prioritize, and coordinate verification of vulnerabilities, with the aim of accelerating vulnerability patching across critical infrastructure and the private sector.</cite>
<cite index="22-8">The effort is housed in the Treasury Department with support from the Pentagon, the Department of Homeland Security (DHS), and the Cybersecurity and Infrastructure Security Agency (CISA).</cite> <cite index="23-1">White House officials said they worked with the Software Engineering Institute (SEI) at Carnegie Mellon University to develop a new platform — the Vulnerability Information and Coordination Environment, or VINTS — to receive third-party reports on AI-discovered vulnerabilities.</cite> <cite index="22-5">VINTS allows for "processing and secure sharing, maintaining validation, prioritization, and disclosure vulnerabilities to ultimately patch and remediate defects and software that underpin essential services."</cite>
Displacing Company-Led Programs
<cite index="10-4,10-5">The Trump administration has taken new steps to assert more control over the rollout of future AI model releases by dictating which companies and entities are allowed access to the latest frontier models — a decision that, until now, was in the hands of American AI giants.</cite> <cite index="4-5,4-6">Anthropic controlled access to its Mythos cybersecurity model through an initiative called Project Glasswing; OpenAI ran a similar programme called Daybreak for its cyber model.</cite> <cite index="1-4">The administration's moves have left the future of company-led initiatives such as Project Glasswing and OpenAI's Daybreak in doubt.</cite>
The shift has already produced measurable consequences. <cite index="17-5">On the evening of June 12, the Commerce Department's Bureau of Industry and Security directed Anthropic to deny all foreign nationals access to its two most capable AI models, Fable 5 and Mythos 5, citing a possible "jailbreak" of the models' cyber guardrails.</cite> <cite index="4-8">The administration blocked Claude Mythos 5 and Fable 5 due to "national security concerns," reinstating access only after weeks of negotiations.</cite> <cite index="10-10">OpenAI last month said it would limit new AI models to "trusted partners" to comply with government requests.</cite>
Tension Between Voluntary Framing and Operational Reality
<cite index="13-13">After a year and a half spent downplaying calls for AI safety regulations, the Trump administration has sharply reversed course, embracing a level of government scrutiny of frontier AI systems before public release — a far stricter stance than the Biden administration took.</cite> A White House official maintained that the government does not "provide approvals for AI releases" and that company participation is "voluntary," yet <cite index="1-5">going forward, these rollouts will require explicit government approval for which partners can be involved,</cite> according to a person familiar with the matter.
<cite index="7-12">Critics warn that centralized control over AI access could stifle innovation, favor politically connected companies, and create a bottleneck that slows American competitiveness against China and other nations.</cite> <cite index="22-6,22-7">Gold Eagle is made possible by the CISA 2015 Act, though a senior official called on Congress to reauthorize the law, which expires in September, warning that "without that reauthorization, this effort is fundamentally challenged."</cite>
Outlook
<cite index="5-6">Companies will be able to submit their most advanced models voluntarily for evaluation, rather than face a blanket testing mandate.</cite> <cite index="12-3">Companies are nonetheless advised to consider whether product release timelines should account for a potential government access period of up to 30 days.</cite> <cite index="2-2,2-3">Gold Eagle is therefore more than a patching program — it is an early test of whether government can coordinate frontier AI for national defence without allowing coordination to become control.</cite>