On August 4th, President Trump hosted AI developers from several major tech companies at the White House to finalize the Executive Office’s voluntary AI framework. This framework intends to have major AI companies voluntarily submit their models to the federal government 30 days before release in order to have them reviewed for cybersecurity purposes and to determine whether they qualify as “covered frontier models”.1 So far, there are only plans to review closed-source AI models – models whose code is not published – but open-source models could enter the conversation as time goes on.2 This framework will allow President Trump to have a more direct say in the development of American artificial intelligence while also following his plans to implement AI models within government agencies.

The implementation of artificial intelligence has been a goal that President Trump has had since his first time holding office from 2016-2020. Back in 2019, Trump passed an Executive Order framing his vision for AI within the United States government and his hopes for continuing the advancement of American AI. 3 With the rapid growth of artificial intelligence, along with his personal ties to the industry, President Trump is now able to fast-track his American AI Action Plan for the United States to take the lead in the AI race and to assert American dominance in the realm of artificial intelligence. 4

All discussions pertaining to the voluntary AI framework have remained exclusively between the administration and the involved companies. 5 The only concrete info known about this framework is the Executive Order passed on June 2nd, in which President Trump outlines his plans to utilize AI for cybersecurity and national defense purposes. He aims to achieve this by working directly with AI developers, allowing them to innovate their models in partnership with the government. 1

How the government plans to review these covered frontier models has not yet been made public and likely will not be, as the Executive has chosen to keep the details of the framework private. Only trusted partners selected by the government and the companies involved will have access to these models once they have been reviewed. It’s also unlikely that the identities of these trusted partners will be made public once chosen, if they haven’t already. 5 The Executive has yet to make any formal statement as to why this benchmarking process and all details regarding it must remain private, making the lack of government transparency under Trump’s presidency a continuing pattern.

There are immense political implications in the administration choosing to keep details of this process private, especially because they are working directly with private tech companies. This privacy leaves American citizens and even other AI companies in the dark about the future of artificial intelligence in the American industry, creating a stir in public opinion regarding the reasons for the administration’s secrecy. As American AI continues to develop and integrate itself into the government, concerns about the democratic state of the country and overall government accountability continue to grow.

Footnotes

  1. The White House, Promoting Advanced Artificial Intelligence innovation and security (opens in a new tab), 2nd June 2026. 2

  2. New York Times, White House Readies A.I. Framework to Review Security Risks (opens in a new tab), 4th August 2026.

  3. Federal Register, Maintaining American Leadership in Artificial Intelligence (opens in a new tab), 11th February 2019.

  4. AI.GOV, AI Action Plan (opens in a new tab), as at 18th August 2026.

  5. Axios, White House plans to keep AI framework under wraps (opens in a new tab), 4th August 2026. 2