Trump signs executive order on frontier AI security and pre-release government access
Executive Order 14409 explicitly rules out mandatory pre-clearance for new models, framing the access as voluntary and contrasting it with the prior administration's approach.
- Government & policy
- Security & misuse
- Era-defining
President Trump signed Executive Order 14409, “Promoting Advanced Artificial Intelligence Innovation and Security,” directing federal agencies to build a framework under which frontier AI developers could voluntarily give the government early access to models before public release. Section 3(b)(ii) specifies developers could provide access to “covered frontier models,” subject to confidentiality and intellectual-property protections, “for a period of up to 30 days before they plan to release such models.”
The order pairs that access framework with cyber-defence measures: it directs the Department of Homeland Security to issue guidance enabling federal agencies, state and local authorities and critical-infrastructure operators to use AI-enabled cybersecurity tools, and establishes a clearinghouse intended to identify vulnerabilities at scale alongside classified benchmarking processes for evaluating advanced models. The order is explicit that none of this creates a legal requirement: Section 3(c) states “nothing in this section shall be construed to authorize the creation of a mandatory governmental licensing, preclearance, or permitting requirement” for developing, publishing or releasing new AI models, including frontier ones.
The administration’s own framing positioned the order as a break from the prior White House’s approach, contrasting it with what officials called the “previous administration’s top-down regulatory approach.” Coverage noted the order nonetheless represented a shift from the administration’s earlier stance of minimal engagement with frontier-model developers, introducing for the first time a formal channel — voluntary in name — through which government could see model capabilities before the public did.
The order’s practical significance became clearer three weeks later, when it supplied the policy basis officials invoked in asking OpenAI to limit the initial release of GPT-5.6, the first known instance of the voluntary framework being used to shape an actual launch decision.