AI-Driven Multimodal Fraud Detection Framework for UID Aadhaar e-KYC Using Biometric and Document Verification

Abstract

Electronic Know Your Customer verification is widely used in banking, government services, and digital platforms to verify user information during digital onboarding. However, most verification methods in use today are based on manual document verification or partially automated verification techniques, which may be time-consuming and prone to identity fraud. This paper proposes an AI-powered multimodal fraud detection framework for Aadhaar Card-based Electronic Know Your Customer verification. The framework is based on integrating document verification, biometric verification, and machine learning for fraud detection. The framework employs Optical Character Recognition technology for verification using user-provided Aadhaar Card images and biometric fingerprint verification for user identification. The extracted user information is further validated through registry cross-checking and One-Time Password email verification. Furthermore, a machine learning algorithm is also used to evaluate the verification process for fraud based on the outcomes of Optical Character Recognition technology, biometric fingerprint verification, and email verification. Additionally, an administrative interface is also proposed for monitoring verification records, fraud patterns, and authentication statistics. The experimental results demonstrate the effectiveness of the proposed framework in ensuring the reliability of verification and reducing fraud in digital identity verification systems.