President Donald Trump is expected to sign an executive order on artificial intelligence and cybersecurity as soon as Thursday.
The draft would require AI developers to provide the US government access to advanced models 90 days before public release and offer early access to critical-infrastructure groups such as banks.
Officials involved in drafting the order include Susie Wiles, Michael Kratsios and Sean Cairncross, and National Economic Council Director Kevin Hassett said the administration is studying a framework similar to the Food and Drug Administration's approval process for drugs.
The Commerce Department has separately announced voluntary pre-deployment testing agreements with Google DeepMind, Microsoft and xAI.
Political pressure for tighter oversight has come from MAGA-aligned figures including Steve Bannon and activist Amy Kremer, who said "AI companies cannot be trusted to 'protect the American people,'" while some tech allies including Marc Andreessen and David Sacks favour voluntary cooperation.
Anthropic has faced growing scrutiny from the Pentagon and federal courts over national-security concerns, and lawmakers have asked Cairncross to help establish a federal process to monitor sudden "frontier AI capability jumps."