Hard Questions, Honest Answers

The proof covers the mapping: under its stated assumptions, no two people produce the same geometric identity. What the proof does not cover — and we don't claim it does — is field performance: matching thresholds, false accept/reject rates, and sensor noise are engineering, measured in deployment, not asserted from math. We say "proven" only where it's proven.

There is no static credential to steal — that's different from claiming nothing can ever be attacked. The shape's protected state never leaves the boundary; raw signals are discarded; and verification tests trajectory, not just position: real identity drifts continuously, a substituted copy has to jump. Presentation attacks and AI mimicry are exactly what the drift/velocity test is designed to catch, and exactly what pilots measure.

Correct — US Patent 11,488,372 (granted 2022) covers the core geometric identity: a provably unique, evolving rendering of an individual. The authentication layer — continuous verification, the tether, instant revocation, AI-agent delegation — is covered in additional patent applications already on file. We keep the two claims separate on purpose.

Passkeys are a real improvement — they end phishing for the login moment. But they verify a device-bound key once, then hand the session to a bearer token, and they federate identity through platform vendors. Ui verifies the living person continuously, revokes the instant continuity breaks, and works identically for humans, organizations, and AI agents. Passkeys fix the front door. Ui removes the doctrine that a moment of verification should buy hours of trust.

Ui is designed around non-retention: raw signals are discarded and only their effect on the geometry persists — no voiceprint, faceprint, or biometric template database. Whether derived geometric state falls under specific biometric statutes is a jurisdiction-by-jurisdiction legal question we treat seriously, with counsel, rather than waving away.

V1 continuous verification is session and context re-validation — the tether model applied to what a scan can capture. Full behavioral-shape verification requires a live signal stream and is the designed roadmap, not the shipped present. We scope this honestly because our customers' auditors will.

Ui V1 supports identity continuity through session and context re-validation. Full behavioral-shape verification requires live signal streams and is part of the pilot and roadmap layer. We are careful not to claim field performance until it is measured in deployment.

Ui is not positioned as a traditional biometric database. Ui is an identity continuity system that can incorporate multiple trust signals, user consent, device continuity, context, behavioral legitimacy, and identity trajectory. Whether derived geometric state falls under specific biometric statutes is a jurisdiction-by-jurisdiction legal question we treat seriously, with counsel.

The geometric mapping model is mathematically defined under stated assumptions. Enterprise security performance — false accepts, false rejects, threshold tuning, sensor noise, device integrity, and live behavioral signal reliability — is validated through implementation, pilot metrics, adversarial testing, and real-world deployment data. We say “proven” only where it is proven.

Recovery is re-verification of the living person — there is no secret to lose, so recovery isn't a reset flow, it's enrollment-grade verification again. Edge cases (injury, major life change, accessibility) are handled by the same principle: the person, verified at higher assurance, is always the root of trust.