White House Keeps AI Cybersecurity Framework Secret from Public

The Trump administration has completed an AI cybersecurity oversight framework stemming from an executive order signed earlier this year. On Tuesday, White House officials invited staff from OpenAI, Anthropic, Google, Meta, Nvidia, and other leading AI companies to receive an overview of the plan. Under the framework, AI developers may voluntarily submit new models to the federal government up to 30 days before public release. The government will then evaluate cyber capabilities using a classified benchmarking system and share results with federal agencies and trusted corporate partners. Open-source models are reportedly excluded. The White House has declined to publicly disclose testing criteria or which models fall under the framework, leaving smaller startups, safety advocates, and independent researchers without key details. Critics warn the secretive approach advantages large incumbent AI companies. 'They're essentially creating an entrenchment program for the big AI model providers,' said one person familiar with the discussions. Brad Carson, president of Americans for Responsible Innovation, stated: 'This is not a handshake deal with tech companies. It's the rulebook for ensuring they don't endanger the public.' A second White House official said the framework is intentionally narrow, focused exclusively on cybersecurity capabilities of the most advanced models, such as Anthropic's Fable and OpenAI's ChatGPT 5.6.