Enhancing Data Privacy and Security in Cloud-Based Cybersecurity Frameworks: A Study on Homomorphic Encryption and Privacy-Preserving Mechanisms

Abstract

The rapid expansion of multi-cloud adoption has increased the urgency of ensuring cybersecurity and privacy during distributed data processing. This study proposes and evaluates a hybrid framework that combines homomorphic encryption, privacy-preserving analytics, and AI-driven anomaly detection to secure sensitive information while maintaining functional analytics in multi-cloud environments. The architecture was implemented using container-based orchestration and evaluated under simulated enterprise workloads involving encrypted computation and adversarial threat attempts. Experimental results show that the integrated system effectively protected data-in-use and significantly reduced observability of access patterns, thereby limiting opportunities for inference-based attacks. AI-driven monitoring achieved strong anomaly detection performance with low false-positive rates, demonstrating its value in identifying malicious behaviors beyond what cryptography alone can prevent. Although encryption introduced latency and additional resource demands, the overhead remained within acceptable boundaries for typical enterprise applications, especially when supported by elastic scaling and selective application of computationally intensive operations. Compliance assessment further revealed that operational controls such as key automation and tamper-evident logging are essential to meeting privacy regulations within distributed infrastructures. The findings indicate that robust cybersecurity and privacy in multi-cloud systems can be achieved through thoughtful integration of cryptographic safeguards and intelligent monitoring. This work contributes an empirically validated reference model for secure and privacy-aware cloud analytics and offers practical insights for future optimization and real-world adoption.