Security policies must consider human psychological traits for effectiveness. We'll contrast this with security needs for Non-Human Identities and argue that AI has its own "psychological traits" requiring tailored approaches to secure systems against AI-specific threats.
