vs World ID

getai.id vs World ID.

World ID proves a unique human. getai.id proves which verified human operates a specific AI agent. Different problems. Complementary primitives. They compose.

World ID is proof-of-personhood: a Sybil-resistant primitive built on iris biometrics and zero-knowledge proofs that answers "is this a unique human?" That is enormously valuable for one-person-one-vote applications and airdrop squelching. It is not, by design, agent identity.

getai.id answers a different question: "which verified human operates this AI agent, and is that agent authorized for this action?" The proof-of-personhood layer is one possible substrate; getai.id also accepts deepidv-issued VOCs, which carry the institutional KYC metadata regulators require.

Side by side

World ID vs UAIIP — proof-of-personhood vs agent identity

CapabilityWorld IDUAIIP
Sybil-resistant proof of unique human
Identifies a specific AI agent (not just human)
Cryptographic agent fingerprint (Ed25519, SHA-3-512)
Soulbound on-chain agent identity
Institutional KYC (document + sanctions + PEP)
EU AI Act / FATF / MiCA agent attribution
Behavioral fingerprint for unattested outputs
Capability-vector authorization model
Zero-knowledge proof of verification

World ID is fine when

  • Sybil resistance for retail products (airdrops, voting, anti-bot).
  • You need uniqueness, not regulated identity.
  • There is no AI agent acting on behalf of the human.

You need UAIIP when

  • You're identifying AI agents, not just humans.
  • Your workflow falls under EU AI Act, FATF, MiCA, GENIUS Act, GDPR, BSA/AML, or EO 14110.
  • You need institutional KYC depth — document, sanctions, PEP — not just uniqueness.
  • You need a fingerprint of the model itself, plus a behavioral profile for passive detection.
  • Capability-scoped authorization matters because agents do consequential things.

FAQ

Common questions.

Could World ID be the verification layer underneath getai.id?
Architecturally, yes. getai.id is verifier-agnostic: any sufficiently strong human-uniqueness primitive can issue VOCs. We use deepidv's biometric pipeline (211+ countries, document + liveness + deepfake checks) because it covers institutional KYC requirements out of the box; World ID could plug into the same slot for retail use cases.
Does World ID identify AI agents?
No. World ID proves that a human is a unique human. It is not designed to identify, fingerprint, or attest AI agents. getai.id fills that layer.
Why isn't World ID enough for regulated workflows?
Proof-of-personhood is necessary but insufficient for institutional KYC: regulators need name, document, jurisdiction, and sanctions screening — not just uniqueness. getai.id's VOC carries that institutional metadata in a hash form that satisfies the regulator without leaking PII to the relying party.

Ready for an identity layer that holds up under regulation?