When AI Can Hack Anything, Identity Becomes Everything

This analysis warns that frontier AI models are making impersonation cheap, scalable, and automated—shifting the primary enterprise attack surface from software vulnerabilities to human identity. It highlights that common authentication controls verify accounts, not the actual human, and that agentic AI and high-fidelity deepfakes enable social-engineering and helpdesk-based breaches at scale, recommending identity verification at the human level (biometrics and auditable human authorization of AI agents) rather than solely detection-based defenses.