FEDERATED AND PRIVACY-PRESERVING AI-BASED PROACTIVE CYBER DEFENSE ARCHITECTURE FOR HEALTHCARE INFORMATION SYSTEMS

The rapid digitalization of Healthcare Information Systems (HIS) has increased exposure to advanced cyber threats targeting sensitive medical data. Traditional signature-based security mechanisms are insufficient against evolving and zero-day attacks. This study proposes a Federated and Privacy-Preserving AI-based cyber defense architecture integrating LSTM-driven anomaly detection, Federated Learning, and Differential Privacy. Experimental evaluation on large-scale healthcare network and IoMT datasets achieved 97% accuracy, 95% recall, and a 0.98 ROC-AUC, with lower false positive rates than conventional IDS systems. The framework provides a scalable, regulation-compliant solution for secure healthcare infrastructures.