Next-Generation Mobile Banking Security: Federated Machine Learning and Post-Quantum Cryptographic Integration for Real-Time Threat Mitigation

The exponential growth of mobile banking has redefined global financial accessibility while simultaneously escalating the sophistication of cyber threats. This study presents a robust, multi-layered cybersecurity framework tailored to the evolving needs of mobile financial systems. The framework integrates three pioneering components: (1) Federated Artificial Intelligence (AI) for privacy-preserving, cross-institutional fraud detection; (2) Post-Quantum Cryptography (PQC) for long-term data integrity against emerging quantum computing threats; and (3) Adaptive Multi-Factor Authentication (A-MFA) leveraging behavioral biometrics for context-aware user verification. Experimental results, drawn from simulated multi-regional banking environments, demonstrate the framework’s efficacy in real-time fraud detection (96.3% accuracy), quantum-resilient encryption (Kyber-768 and Dilithium-3 with negligible latency), and enhanced user authentication (98.1% success rate, 1.4% FRR). Furthermore, the federated learning model maintained high accuracy under non-IID conditions, underscoring its robustness in heterogeneous financial networks. This research fills critical implementation gaps in the literature and provides a scalable, regulation compliant model for secure mobile banking in the post-quantum era. The findings advocate for a proactive shift toward integrated, intelligent, and future-proof cyber defense mechanisms within the global financial ecosystem.