White House Expands Oversight of Frontier AI Models

The Trump administration has moved to increase its influence over which organizations receive early access to the most advanced artificial-intelligence systems from companies such as OpenAI and Anthropic. The newly announced “Gold Eagle” cybersecurity initiative is positioned as a clearinghouse through which the White House can evaluate and coordinate access to frontier models, marking a shift from the prior practice in which AI firms alone controlled early-partner programs.

BenzingaWhite House Now Controls Access to Frontier AI ModelsAssociated Press

Historically, OpenAI’s “Daybreak” and Anthropic’s “Project Glasswire” have selected trusted enterprises, government agencies, and research groups for pre-release testing of their latest models. In June, President Donald Trump issued an executive order urging frontier AI developers to voluntarily share early versions for security testing. Earlier this year, the administration asked OpenAI to limit distribution of its GPT-5.6 model and temporarily blocked Anthropic’s Claude Mythos 5 and Fable 5 on national-security grounds before restoring access after negotiations. The push coincides with heightened competition from China, where startups such as Moonshot AI have released models that approach the performance of U.S. offerings.

BenzingaWhite House Now Controls Access to Frontier AI ModelsAssociated Press

• President Donald Trump – issued the June executive order on AI security.

Benzinga

• OpenAI – developer of the GPT-5.6 model and the Daybreak early-access program.

White House Now Controls Access to Frontier AI Models

• Anthropic – creator of Claude Mythos 5 and Fable 5, operating Project Glasswire.

White House Now Controls Access to Frontier AI Models

• White House official (anonymous) – provides the administration’s public comments on the program.

White House Now Controls Access to Frontier AI Models

• David Sacks – former White House AI czar, now a vocal critic of the administration’s approach.

Benzinga

The White House emphasizes that its involvement is collaborative and voluntary. A senior official told CNBC that “the Administration continues to collaborate with all of America’s frontier labs to strengthen the security of this technology without stifling innovation.” The same official clarified that “decisions on timing and scope of releases rest entirely with the companies,” and that the administration does not grant permissions for AI product launches. The administration’s rationale is to balance national-security concerns with maintaining U.S. leadership in AI amid rapid advances from foreign competitors.

White House Now Controls Access to Frontier AI Models

Former AI czar David Sacks warned that the United States could fall behind, stating, “This is how you lose the AI race.” He added, “The rest of the world won’t play by our rules if we bog ourselves down,” reflecting industry unease that government oversight may hinder rapid innovation and disrupt existing private-sector early-access frameworks.

White House Now Controls Access to Frontier AI Models

Industry insiders note that the new oversight has introduced uncertainty for programs such as Anthropic’s Project Glasswire and OpenAI’s Daybreak. After weeks of negotiations, Anthropic’s models were released, but the episode underscored the administration’s willingness to intervene before a model reaches the market.

White House Now Controls Access to Frontier AI Models

Sources agree that the Gold Eagle program is now operational, yet they differ on the extent of its authority. While one White House official asserts that all decisions remain with the companies, another anonymous source emphasizes that the administration “does not grant permissions for AI product launches.” No public criteria have been disclosed for determining which entities qualify for early access, leaving the precise scope of government control unclear.

White House Now Controls Access to Frontier AI Models

• “the Administration continues to collaborate with all of America's frontier labs to strengthen the security of this technology without stifling innovation.” — White House official

White House Now Controls Access to Frontier AI Models

• “decisions on timing and scope of releases rest entirely with the companies.” — White House official

White House Now Controls Access to Frontier AI Models

• “ He wrote, "This is how you lose the AI race.” — David Sacks, former White House AI czar

White House Now Controls Access to Frontier AI Models

• “This is how you lose the AI race. The rest of the world won't play by our rules if we bog ourselves down.” — David Sacks

White House Now Controls Access to Frontier AI Models

• “A White House official, speaking on condition of anonymity, stated that the administration does not grant permissions for AI product launches by private firms.” — Anonymous White House official

White House Now Controls Access to Frontier AI Models