“The White House's AI safety directive is being developed following Anthropic's revelation of its Mythos model, which is considered a potential global cybersecurity risk.”