WHITE HOUSE CUTTING A DEAL WITH BIG AI: VOLUNTARY REVIEW STANDARDS IN ADVANCED TALKS WITH OPENAI, GOOGLE, AND ANTHROPIC The Trump White House is in the final stages of negotiating a voluntary AI pre-release review framework directly with OpenAI, Google, and Anthropic, with an announcement possible as early as next week. The framework builds on the executive order signed in June that asked AI companies to share their most powerful new models with the government up to 30 days before public release, and would formalize what both sides have already been doing informally since the GPT-5.6 review in late June. The deal is voluntary in name but mandatory in practice. Both OpenAI and Anthropic have already committed to briefing the government before future frontier model releases. What is being negotiated now are the specific terms: how long the government gets to review, which agencies are involved, what the classified benchmarking process looks like, and whether companies face any consequences if a review raises concerns. The backdrop here is significant. The government delayed GPT-5.6’s public launch. It reinstated export controls on Claude Fable 5 for nearly three weeks before lifting them. The White House has demonstrated it will act on these reviews. Now it wants the process codified, and the AI labs appear willing to put their names on it. What gets written into this framework will shape how every major frontier model reaches the public for years to come. Keywords: White House AI release standards, voluntary AI framework, frontier model review, OpenAI government deal