Machine Learning in Detecting Insider Threats Within Organisations

Abstract

The Insider threats have emerged as one of the most critical challenges in contemporary cybersecurity. Recent studies indicate that approximately 25% of cyberattacks originate from insiders, with these incidents often resulting in greater financial and operational damages [3]. This paper investigates the application of machine learning (ML) techniques for the detection and prediction of insider threats within organizations. We review the classification of insider threat types—including data theft, privilege abuse, privilege escalation, and sabotage—and examine both heuristic and ML-based detection methods. The proposed framework integrates supervised learning, unsupervised clustering, and deep learning methods to enhance detection accuracy and minimize false positives. Our results, drawn from recent advances in ML applications, underscore the potential for robust, real-time threat detection solutions that address both known and emerging insider attack vectors.