AI-Driven Fraud Detection and Cybersecurity Framework for Protecting U.S. Healthcare Systems Against Emerging Cyber Threats

Abstract

The rapid digitalization of the U.S. healthcare sector has created a large and increasingly valuable attack surface that cybercriminals continue to exploit. This paper examines two major threats facing healthcare organizations: financial fraud involving fraudulent claims and billing, and cybersecurity attacks such as ransomware and data breaches. It proposes an integrated, AI-driven framework for addressing both threats. Drawing on peer-reviewed literature, the paper first examines the changing healthcare threat landscape, including the sharp increase in hacking-related breaches and the operational and patient-safety consequences of ransomware. It then reviews artificial intelligence and machine learning approaches to fraud detection, including supervised and unsupervised classifiers, deep-learning anomaly detection, and natural language and network analytics. These capabilities are then considered within established cybersecurity frameworks, particularly the NIST Cybersecurity Framework and HIPAA. The paper's primary contribution is a layered framework that connects AI-based fraud detection with the Identify, Protect, Detect, Respond, and Recover functions of cybersecurity risk management while incorporating privacy-preserving techniques and human oversight. The paper concludes by discussing ongoing challenges involving model interpretability, data privacy, class imbalance, and benchmarking gaps, as well as future opportunities such as real-time prevention, federated learning, and explainable AI.