White House National Economic Council Director Kevin Hassett said on Monday that the Trump administration does not intend to establish a new federal agency to oversee advanced artificial intelligence models.
Speaking to CNBC, Hassett described a strategy centered on cooperation with private companies to prevent AI systems from causing significant harm prior to being made public. He identified Treasury Secretary Scott Bessent, National Cyber Director Sean Cairncross and White House Chief of Staff Susie Wiles as the principal officials coordinating those efforts with industry representatives.
The comments follow a Bloomberg report earlier that outlined a draft executive order that could alter current cybersecurity information-sharing programs to bring AI companies into their scope and address risks tied to emerging technologies. Hassett confirmed that the report had prompted discussions but said President Donald Trump had not yet chosen whether to issue an executive order or to rely on the existing oversight framework.
Hassett also said that meaningful progress has been made in recent weeks through dialogue between government and private-sector participants. He framed the work as an ongoing process of coordination, rather than the creation of new institutional structures.
Context and next steps
- Coordination approach: Administration officials are engaging with private-sector actors to identify and limit potential harms from advanced AI models before public deployment.
- Possible policy tool: A draft executive order has been reported that could expand existing cybersecurity information-sharing mechanisms to include AI firms, but no final decision has been made on issuing such an order.
- Leadership: Treasury Secretary Scott Bessent, National Cyber Director Sean Cairncross and White House Chief of Staff Susie Wiles are leading interagency and private-sector coordination.
Summary
The administration has rejected the idea of creating a new agency to oversee advanced AI models, favoring a collaborative model with private companies and potential adjustments to existing information-sharing programs. Officials say the president has not yet decided whether to formalize those adjustments through an executive order. Government and industry representatives report recent progress through ongoing conversations.
Key points
- No new federal agency will be created to police advanced AI models, according to the White House National Economic Council Director.
- Coordination is being led by top administration officials working with private-sector companies to mitigate harmful outputs from AI models before public release.
- A draft executive order has been reported that could modify existing cybersecurity information-sharing programs to include AI companies; the president has not decided whether to proceed.
Risks and uncertainties
- Uncertainty over whether an executive order will be issued leaves the exact legal or regulatory changes unresolved - affecting how quickly any new requirements would take effect.
- Relying on coordination with private firms rather than creating a new agency could leave gaps in oversight depending on the scope and enforcement of voluntary or modified information-sharing programs.
All assertions above reflect statements and developments described by administration officials and reported drafts; they do not introduce additional facts beyond those provided by those sources.