Financial regulators worldwide are monitoring Anthropic's Mythos AI model after banking executives raised concerns about its potential to facilitate cyberattacks. The model is also reportedly being used by the US National Security Agency.