Left-right coalition seeks release of White House AI framework

A coalition of over two dozen organizations spanning both left and right wings is calling on the White House to make its voluntary AI security framework publicly available, according to a letter obtained first by Semafor. The letter, penned by Americans for Responsible Innovation and the Center for Democracy and Technology, emphasizes the government's duty to outline the process for assessing frontier AI models before they are widely deployed.

The White House first revealed the much-anticipated framework to a limited group of tech companies last month. Since then, the pressure to disclose the framework has grown, following incidents such as OpenAI's assault on AI model repository Hugging Face and the release of OpenAI's latest model, Astra. Protect Democracy, a nonprofit, recently filed a lawsuit against the Trump administration to secure a Freedom of Information Act request on the framework, which was part of Trump's June AI order.

The White House met its internal deadline of August 1 to publish the voluntary framework for reviewing certain powerful AI models before their public release, but only a select few have actually had access to it. As public opinion increasingly leans towards AI safety regulations and opposes data center growth, the administration is expected to face more calls for transparency regarding the framework's contents.

The letter highlights that a lack of accountability regarding the governance of a rapidly evolving technology could undermine democratic values and erode public trust. Furthermore, the groups argue that transparency is crucial to prevent the dominant AI labs from monopolizing the most advanced models while leaving open-source companies in the dark. They also assert that Congress should be included in the decision-making process.