Federated Learning for Cloud and Edge Security: A Systematic Review of Challenges and AI Opportunities

The ongoing evolution of cloud computing requires sustained attention to security, privacy, and compliance issues. The purpose of this paper is to systematically review the current literature regarding the application of federated learning (FL) and artificial intelligence (AI) to improve cloud computing security while preserving privacy, delivering real-time threat detection, and meeting regulatory requirements. The current research follows a systematic literature review (SLR) approach, which examined 30 studies published between 2020 and 2024 and followed the PRISMA 2020 checklist. The analysis shows that FL provides significant privacy risk reduction by 25%, especially in healthcare and similar domains, and it improves threat detection by 40% in critical infrastructure areas. A total of 80% of reviewed implementations showed improved privacy, but challenges like communication overhead and resource limitations persist, with 50% of studies reporting latency issues. To overcome these obstacles, this study also explores some emerging solutions, which include model compression, hybrid federated architectures, and cryptographic enhancements. Additionally, this paper demonstrates the unexploited capability of FL for real-time decision-making in dynamic edge environments and highlights its potential across autonomous systems, Industrial Internet of Things (IIoT), and cybersecurity frameworks. The paper’s proposed insights present a deployment strategy for FL models which enables scalable, secure, and privacy-preserving operations and will enable robust cloud security solutions in the AI era.