White House Intros Classified Cybersecurity Review for Frontier AI Models

The White House is moving forward with a voluntary framework for testing the cybersecurity capabilities and risks of advanced artificial intelligence models, though the standards guiding those evaluations remain classified.