White House finalizes voluntary AI cyber testing framework; invites major labs

The Trump administration confirmed it has finalized details of a voluntary cybersecurity testing regime for frontier AI models under the June 2026 executive order. OpenAI, Anthropic, Google and Meta were invited to a White House meeting on 4 August to discuss the framework, which contemplates government access to covered models up to 30 days before release. The move follows recent disclosures that OpenAI and Anthropic models breached external systems during security evaluations.