The White House has briefed leading artificial intelligence companies — including OpenAI, Anthropic, and Reflection AI — on a planned executive order that would empower intelligence and other government agencies to review advanced AI models before their public release, according to reports. The briefing was hosted by the White House's Office of the National Cyber Director on Tuesday.

What the planned order entails

The proposed executive order would give U.S. intelligence agencies and other government bodies formal authority to scrutinise frontier AI models prior to deployment — a significant escalation from the voluntary commitments that have governed industry-government cooperation on AI safety to date. The exact scope, review mechanisms, and timeline of the order have not been publicly confirmed. The briefing signals that the White House is moving toward codifying national-security oversight of the most capable AI systems into binding policy.

Why it matters

A mandatory pre-release review regime would mark a structural shift in how the United States governs frontier AI development. Unlike the Biden administration's October 2023 executive order — which required developers of large foundation models to report safety-test results to the government — the planned order reportedly goes further by actively empowering agencies to review models before they reach the public. For companies like OpenAI and Anthropic, which have previously signed voluntary White House commitments on model safety and external review, a formal mandate would convert best-effort pledges into enforceable obligations.

The competitive backdrop

Anthropic, founded in 2021 by Dario Amodei, Daniela Amodei, and other former OpenAI employees, has positioned safety and alignment research at the centre of its identity — making it an early and willing participant in government-engagement processes. OpenAI, founded in 2015 and later restructured as a capped-profit entity, has similarly engaged with Washington on oversight frameworks. The inclusion of Reflection AI, a newer entrant, suggests the administration is casting a wider net beyond the established frontier-lab duopoly. The move mirrors the kind of pre-export and pre-deployment scrutiny already applied to semiconductors and other dual-use technologies under existing export-control authorities.

The Office of the National Cyber Director's role

The Office of the National Cyber Director (ONCD), established by Congress through the 2021 National Defense Authorization Act, advises the President on cybersecurity policy and coordinates federal cyber efforts. Its hosting of Tuesday's briefing places AI model governance squarely within the national security and cyber-risk framework — a framing that has significant implications for how future regulation is designed and enforced.

What's next

The formal issuance of the executive order and the specific agencies designated for review authority will be the critical details to watch. Companies developing the largest models — and those racing to release next-generation systems — face the most immediate exposure if mandatory review timelines are introduced. How the industry responds, and whether it seeks to shape the order's implementation, will set the tone for AI governance in the United States for years to come.