Decentralized KYC Verification Systems with AI Facial Recognition

Authors

  • Dr. Richard Collins Faculty of Artificial Intelligence, University of Dublin Global, Ireland Author

DOI:

https://doi.org/10.63345/

Keywords:

Decentralized Identity, KYC/AML, Verifiable Credentials, Decentralized Identifiers, Facial Recognition, Liveness Detection, Zero-Knowledge Proofs, Privacy, ISO/IEC 30107-3, NIST SP 800-63-3

Abstract

Know-Your-Customer (KYC) processes are foundational to anti–money laundering (AML) compliance but are often centralized, costly, and privacy-invasive. This manuscript proposes and analyzes a decentralized KYC (dKYC) architecture that binds government-issued identity attributes to a user-held digital wallet using Decentralized Identifiers (DIDs), Verifiable Credentials (VCs), and AI-based facial recognition with liveness detection. The design minimizes data movement and disclosure by leveraging selective-disclosure credentials and zero-knowledge proofs (ZKPs) so that relying parties can verify eligibility (e.g., age, residency, risk tier) without accessing raw personally identifiable information (PII). We map the system to prevailing assurance frameworks (FATF digital ID guidance; NIST SP 800-63-3 Identity Assurance Levels), biometric performance/testing standards (ISO/IEC 30107-3 PAD; ISO/IEC 19795-1), and emerging regulatory regimes (EU AI Act, eIDAS 2.0, RBI directives on KYC and V-CIP). A comparative evaluation (simulated) contrasts centralized KYC with the proposed dKYC across error rates, onboarding time, and privacy risk. Results suggest dKYC can reduce median onboarding time by ~35–55%, lower document leakage risk through off-chain storage and selective disclosure, and maintain biometric security with PAD Level 2 while controlling demographic differential performance via auditable testing and thresholding. The paper contributes (i) a standards-aligned reference architecture; (ii) an evaluation plan referencing recognized biometric metrics (FMR/FNMR; APCER/BPCER); and (iii) a regulatory and governance blueprint for cross-jurisdictional adoption. Anchors to standards and regulation: DIDs and VCs enable user-controlled identifiers and machine-verifiable credentials; FATF clarifies when digital ID can satisfy CDD; NIST provides assurance and authentication guidance; ISO standards define biometric testing and anti-spoofing; the EU AI Act and eIDAS 2.0 frame lawful high-risk biometric use and wallets; RBI updates enable video-based KYC in India.

Downloads

Download data is not yet available.

Published

02-04-2026

Issue

Section

Original Research Articles

How to Cite

Decentralized KYC Verification Systems with AI Facial Recognition. (2026). Scientific Journal of Artificial Intelligence and Blockchain Technologies, 3(2), Apr(01-10). https://doi.org/10.63345/

Similar Articles

21-30 of 92

You may also start an advanced similarity search for this article.