White House Plans Executive Order Requiring AI Model Review Before Release
WASHINGTON — The White House is developing an executive order that would require government review of advanced AI models before public release, according to The Decoder.
Senior administration officials have briefed executives from Anthropic, Google, and OpenAI on the proposed review process, which would establish a formal federal evaluation framework for frontier AI systems before they reach consumers, the report said. The proposed order would represent a departure from the administration’s approach over the past year, which focused on reducing regulatory barriers for the AI sector, according to the report.
The initiative was reportedly triggered by Anthropic’s forthcoming “Mythos” model, which appears to have raised concern within the administration, according to the report. Details of what specific capabilities prompted the response were not disclosed in the reporting.
If enacted, the executive order would establish formal federal oversight of AI deployment, potentially reshaping release timelines for the nation’s three largest AI developers. The companies have collectively invested tens of billions of dollars in model development and have operated under relatively permissive regulatory conditions over the past year.
The proposed review process would affect an industry that has moved rapidly to deploy increasingly capable systems. OpenAI, Google DeepMind, and Anthropic have each released multiple model generations in recent months, with competitive pressure driving accelerated deployment schedules.
The briefing of all three major labs suggests the administration is seeking industry input before finalizing the order’s terms, a pattern consistent with previous technology policy initiatives. However, the scope and timeline of the proposed review process remain unclear.
For enterprise customers and developers building on these platforms, a pre-release government review could introduce delays between model completion and commercial availability — a consideration for companies that have built business strategies around rapid access to new AI capabilities.
The policy shift comes as governments worldwide grapple with how to balance innovation incentives against potential risks from increasingly powerful AI systems. The European Union’s AI Act enforcement deadline approaches in August 2026, and several US states have advanced their own AI governance frameworks in the absence of comprehensive federal legislation.
Neither the White House nor the three AI companies have publicly commented on the reported briefings.
