Quantum-Secure AI Models for Next Generation Network Cybersecurity: Architecture, Implementation, and Case Study in Cameroon

Abstract The rapid proliferation of Next Generation Networks (NGN), encompassing 5G, edge computing, cloud-native infrastructures, and massive Internet of Things (IoT) deployments, has dramatically expanded the attack surface for cyber adversaries. Concurrently, the advent of cryptographically relevant quantum computers threatens to render widely deployed public-key cryptosystems such as RSA and Elliptic Curve Cryptography (ECC) obsolete through algorithms like Shor’s factoring algorithm. This dual challenge necessitates a paradigm shift toward security architectures that are simultaneously resilient to classical and quantum-era threats. This paper proposes the Quantum-Secure AI (QSAI) framework, a novel cybersecurity architecture that integrates NIST-standardized Post-Quantum Cryptography (PQC) specifically ML-KEM (CRYSTALS-Kyber) and ML-DSA (CRYSTALS-Dilithium) with federated learning (FL) for privacy-preserving distributed intelligence and AI-driven intrusion detection with adversarial robustness. The QSAI framework operates across four interdependent layers: a Device/Edge Layer, a Federated Learning Layer with quantum-safe aggregation, an AI Defense Layer incorporating adversarial training and quantum-inspired optimization, and a Security Orchestration Layer employing zero-trust principles and reinforcement learning-based policy engines. Experimental evaluation on benchmark datasets (CICIDS2017, UNSW-NB15) demonstrates that the centralized baseline achieves 98.1% detection accuracy, while the federated QSAI variant attains 93.3% accuracy with the critical advantage of preserving data privacy and quantum-resistant confidentiality, a 16.9% improvement over federated models under adversarial poisoning conditions. PQC integration introduces minimal overhead (approximately 11–14% additional latency per cryptographic operation). A dedicated case study examines deployment feasibility within Cameroon’s evolving NGN ecosystem, addressing local infrastructure constraints, the 156% increase in cyberattacks observed between 2020 and 2023, and strategic recommendations for quantum-safe adoption in resource-constrained African telecommunications environments. Keywords: Post-Quantum Cryptography, Federated Learning, Intrusion Detection Systems, Next Generation Networks, 5G Security, CRYSTALS-Kyber, Adversarial Machine Learning, Zero-Trust Architecture, Cameroon NGN, Quantum Computing Threats