Anthropic PBC is looking into reports of unauthorized access to its Claude Mythos AI model, which is reportedly capable of facilitating dangerous cyberattacks. This incident occurred shortly after the company's announcement of a limited testing release of the model.