India’s cybersecurity agency, CERT-In, has gained access to Anthropic’s Claude Mythos AI model. CERT-In is now using this model to test the security of its systems as AI models become better at identifying cyber weaknesses.