Trump administration’s AI pre-review system in final stages… OpenAI, Anthropic and Google submit draft revisions

The Trump administration is pushing a new framework that would let the government review frontier AI models for up to 30 days before public release.

The White House cyber office shared a draft with OpenAI, Anthropic, and Google, and the companies have submitted revisions.

The plan builds on a June AI executive order, but key issues remain unresolved, including how to define frontier models and whether open-source systems are covered.

If adopted, the policy could set a major precedent for AI safety and national security oversight across the industry.