White House framework for testing new AI models remains hidden as concerns mount
The White House has not publicly released the voluntary framework for testing frontier AI models that it finalized in August.
Washington - The White House has not yet disclosed the voluntary framework they finalized for testing advanced AI models, despite concerns raised by current and former employees of AI companies. Anthropic, the company behind the Claude model, disclosed that they terminated several research projects using Claude due to potential risks of developing biological weapons.
The White House's framework, established in August following President Trump's executive order in June, aimed to enhance model security by allowing AI firms to share their most advanced models with the federal government 30 days before release. However, the criteria for evaluation remain undisclosed, leaving it unclear what measures companies must meet or if they are revealing new breakthroughs.
Participation in the review process and results are also not mandated. Sam Altman, CEO of OpenAI, expressed support for the review process, emphasizing its significance as models advance. Anthropic researcher Jacob Coxon and former OpenAI employee Evan Hubinger resigned, expressing deep concerns about AI potentially threatening humanity.
The Center for Democracy and Technology urged the White House to release the framework, stating the public's right to know the government's interpretation of the law. Despite warnings, the administration continues to support AI development to maintain global dominance over China.
Written by urgent.news from CBS News's reporting — not their text. Machine-written — may contain errors; check the original before relying on it.
