Trump's voluntary AI testing framework excludes open models

Trump administration’s voluntary AI testing framework reportedly excludes open models and will not restrict them after release, raising concerns about oversight. Axios says guidelines allow a 30-day government review period but only apply to closed-source models with unnamed state-of-the-art capabilities and national security risks. Companies including OpenAI, Anthropic, and Google attended a White House briefing; details remain undisclosed. The framework stems from Trump’s June executive order requesting pre-release sharing. It is voluntary, but frontier labs seek guidance; ambiguous terms could hurt smaller providers.