The Role of Explainable Artificial Intelligence in Diagnosing and Mitigating Potential Data Breach Risks in Healthcare

Abstract

In this study, the impact of Explainable Artificial Intelligence (XAI) approaches on the source detection of potential data breaches and their role in mitigating risks are investigated. The input features of datasets containing cyber-attacks and potential threats were analyzed and their potential to detect threats was evaluated. It is known which types of cyber-attacks are realized with which input features in these datasets. As a result of the evaluations obtained, it is seen that the most needed input features in mitigating attacks on healthcare services are the attacker's targets, vulnerable network assets information, and violation points of security policies. As a result, it is predicted that XAI applications can play an effective role in reducing potential risks and data leaks in patient care applications by interpreting how dataset characteristics affect decisions in artificial intelligence models.