An AI-Driven and Risk-Aware Digital Identity Protection Framework for Secure IoMT Environments

With the rapid expansion of the Internet of Medical Things (IoMT), the importance of digital identity–based security has significantly increased. However, conventional static authentication mechanisms are insufficient to effectively address various identity misuse and abuse attacks. In this study, we model digital identity as a dynamic security entity and propose an AI-based framework that integrates a risk scoring model—combining unsupervised anomaly detection with context-aware analysis—and a multi-level risk-adaptive access control mechanism (Permit, Step-Up, Restrict). Experimental results using an extended version of the CERT Insider Threat Dataset tailored for IoMT environments provide proof-of-concept evidence that the proposed method can achieve an AUC of approximately 0.927, orange demonstrating effective discrimination between normal and malicious behavioral patterns. Furthermore, the framework maintains a low False Restriction Rate of around 1.605% while still detecting attacks at a meaningful level, thereby achieving a balance between security and usability. This study highlights the feasibility of a risk-adaptive digital identity protection framework that dynamically evaluates digital identity and adaptively responds based on risk levels in IoMT environments.