AI Agents Don’t Have Identities and That’s a Security Crisis

Only 21.9% of organizations treat AI agents as independent identities. The rest use shared API keys and service accounts. Current identity models — built for humans, services, and bots — don't fit autonomous agents. The IETF, NIST, SPIFFE, and Google's A2A protocol are building the pieces, but nobody has assembled the full stack yet. Here's the five-layer identity architecture AI agents need, and what you can do today before the standards land.