IJAEMS
IJAEMS
Cloud computing systems support scalable data storage, application hosting, and digital services, but their distributed nature exposes them to malware, intrusion, data leakage, denial-of-service attacks, and unauthorized access. In this research, authors have suggested an innovative AI-based security architecture that can classify threats and respond to them automatically. The architecture features cloud traffic monitoring, security-log preprocessing, feature extraction, machine-learning classification, risk assessment, and automated incident response. AI analyzes data on the network and user and application activity to detect patterns and categorize threats by type and severity. A response engine automatically acts on appropriate actions, such as blocking malicious traffic, isolating compromised resources, limiting suspicious accounts, creating alerts, etc., and beginning recovery operations. By enabling the classification model to adjust to new attack trends and increase the accuracy of detection, continuous feedback can be provided. The proposed architecture contributes to decreasing the manual effort required for security, decreasing the delay in responding to a security event, preventing false alarms, mitigating securityrelated operational risk, and enhancing scalability, visibility, and protection in cloud infrastructures. The proposed AI-based security architecture achieved 96.40% classification accuracy, demonstrating reliable automated threat identification across multiple cloud attack categories. It offers a smart and predictive security solution to today's cloud computing systems.