Secure-Pay: Smart AI Defense for UPI And Cyber Threats

Abstract

Due to growing advanced cyber attacks, the security systems used to monitor payment fraud should be smartly tuned in real-time to identify such threats immediately. Our paper proposes an AI-Driven Multi-Threat Cybersecurity & UPI Fraud Detection System that unites elements within a single ecosystem to detect browser extensions aimed at, ransomware activities, and UPI/QR code-related financial fraud. The deep-learning architectures such as LSTM, GRU, Autoencoders, and Graph Neural Networks help in the detection faculties of the system which are hidden and attack vectors that traditional rule-based methods could not trace. The Behavioral Monitoring System checks extension behaviors and tracks file actions, user activities, and digital payment trends to determine irregularities, such as during pre-encryption ransomware activity or during altered QR payment. To top that up, the framework adds layers for fraud verification such as device fingerprinting, transaction scrutiny, and AI-based screenshot verification. The proposed mechanism proposes a scalable path, however remains a potent option, to counter contemporary cybersecurity and fintech demands with real-time alerting, light deployment, and high accuracy combined with continuous learning capabilities. Experimental results verify its success in ensuring proactive threat mitigation across diverse attack vectors while maintaining minimum false positives.