AI Under Siege: Building Cyber-Resilience Across Cloud, Edge, and Autonomous Systems

Abstract: As AI systems become increasingly embedded across cloud infrastructures, edge devices, and autonomous platforms, their exposure to sophisticated cyber threats is rapidly intensifying. This research presents a comprehensive framework for building cyber-resilient AI systems capable of withstanding, detecting, and adapting to adversarial attacks across distributed environments. By integrating adversarial training, zero-trust edge gateways, federated anomaly detection, and blockchain-secured model updates, the proposed architecture enhances security without compromising system performance. Experimental simulations across diverse threat scenarios demonstrate significant improvements in accuracy under attack, faster recovery times, and higher overall resilience scores compared to baseline models. The findings offer a strategic blueprint for developing AI systems that are not only intelligent but inherently secure, adaptable, and robust in dynamic threat landscapes. Keywords: Cyber-resilience, Artificial Intelligence, Cloud security, Edge computing, Autonomous systems, Adversarial robustness, Zero-trust architecture, Federated learning, Blockchain, AI security