White House finalizes secretive AI review framework, smaller labs object

The White House has privately finalized a voluntary framework letting U.S. officials review top AI models before release, but is keeping details secret, drawing criticism from smaller labs. Only a handful of firms attended Tuesday's briefing. The framework covers closed-source state-of-the-art models posing national-security risks, allowing 30-day government review. Open-weight models appear excluded, and key definitions remain undefined. Critics, including lawmakers and policy groups, say secrecy undermines accountability and may entrench government-approved closed-source models while excluding open-weight systems Washington is worried about.