Human Principal (proof-of-human root).
A human principal is the verified natural person at the root of an AI agent's delegation chain. Proof-of-human attestations and identifiers let a site confirm that a real, live person stands behind an agent's action.
Most agent security work so far answers "which agent is this?" A human principal answers the question underneath: "is there a real person behind it, and which one?" Sites want this for per-person rate limits, abuse-protected free tiers, and accountability when an agent misbehaves.
The variants differ in what they reveal. A VHRA as drafted proves a liveness event happened for an identified person, without shipping biometrics to the verifier. World ID style proofs aim to show a unique human without revealing who. A selfie-derived persistent token, such as Unico's Human Token, recognizes the same face across services. PYMNTS's "Know Your Human" (February 27, 2026) frames the business practice: confirm a real person authorized the instruction and that the agent stayed within it. Sources: draft-zagarella-verified-human-root-01, World's announcement with Okta, and IDTechWire on Unico Human Token.
Our reading. Pick the weakest proof that solves the problem. Uniqueness without identity fits abuse control; identified proof fits regulated payments, where KYC already applies. In Trusted Agent Delegation terms, this is the Unique tier, and it carries the highest privacy cost. Pair it with Know Your Agent checks; neither replaces the other.
Common questions
What is a human principal for an AI agent?
The human principal is the real person on whose authority an AI agent acts. Proving a human principal means showing that a verified, live person sits at the root of the agent's delegation chain, not only an account or an API key.
What is a Verified Human Root Attestation (VHRA)?
VHRA is defined in the individual IETF draft draft-zagarella-verified-human-root-01 (August 24, 2026). It is a signed attestation that a biometric proof-of-human check of an identified person happened at a specific issuance event. A delegation chain binds it at its root, and audit records can reference it, so an agent action traces to a person without the verifier receiving biometric data.
Is Human Principal an Auth0 product?
The name comes from Okta, not Auth0. World's April 17, 2026 announcement said Okta is planning a product called Human Principal, with World ID as one of its first integration partners, and Okta runs a design-partner beta sign-up. World's post notes that future products are not commitments. Check Okta for current availability.
Does proof of human prove the person approved a specific action?
No. A proof-of-human root shows a live person stands behind the agent. It does not show that the person approved this purchase or this data access. That still needs consent and authorization records, such as scoped OAuth grants or signed payment mandates.
Related terms
In the guides
Biometric Authentication: A Practitioner's Guide to Fingerprint, Face, and Beyond
How modern biometric authentication actually works: device-local templates, signed assertions, liveness detection, and where the privacy story is real vs marketing.
Identity Verification and Proofing (IDV/KYC): A CIAM Guide for 2026
How to prove a real person matches a claimed identity at signup: document capture, liveness, authoritative-data checks. The 2026 stack, the deepfake escalation, and where CIAM ends.
Know Your Agent: Identity for AI Agents That Buy on a Customer's Behalf
What a CIAM team must build when a customer's AI agent shows up at login, account, or checkout: delegated consent, agent sessions, step-up, audit, and revocation.