White House meets OpenAI, Anthropic, Google, Meta on voluntary AI cyber testing framework

The Trump administration held a staff-level White House meeting with representatives from OpenAI, Anthropic, Google, Meta and Nvidia to discuss the finalized voluntary cybersecurity testing framework for frontier AI models. Officials indicated the tests would not apply to open-weight models. The framework stems from a June executive order and follows recent disclosures of model-related security incidents. The process remains unpublished.