ADVANCING THREAT DETECTION THROUGH ARTIFICIAL INTELLIGENCE AND MACHINE LEARNING ENHANCED CYBERSECURITY AUDITS

Abstract

The increasing complexity of cyber threats and the intensification of regulatory demands have elevated cybersecurity auditing into a strategic imperative for organizations across sectors. Traditional audits, often reliant on manual verification and rule-based models, have shown limitations in efficiency, scalability, and accuracy. This study explores how artificial intelligence (AI) and machine learning (ML) transform cybersecurity auditing by enhancing compliance automation, threat detection, auditor trust, continuous monitoring, and sector-specific assurance in finance and healthcare. Seven hypotheses were developed, grounded in computational learning theory, risk governance, and explainable AI frameworks, to examine both the direct and moderating effects of these technologies on audit effectiveness. A quantitative cross-sectional survey was conducted with 245 professionals, including auditors, compliance officers, IT managers, and security executives from finance, healthcare, energy, and government organizations. Data were analyzed using multiple regression and structural equation modeling to assess hypothesized relationships. The findings provide strong empirical support: AI-based compliance automation significantly improved audit efficiency (β = 0.42, p < .001), reducing cycle times and minimizing human error; ML-driven models enhanced threat detection accuracy (β = 0.47, p < .001), lowering false positives and identifying complex anomalies; and explainable AI features increased auditor trust in automated outcomes (β = 0.36, p < .001), particularly among less experienced professionals. Continuous auditing enabled by AI and robotic process automation was strongly associated with reduced organizational risk exposure (β = −0.51, p < .001), demonstrating tangible governance benefits. Domain-specific analysis confirmed that AI integration improved fraud detection rates in financial services (β = 0.44, p < .001) and strengthened HIPAA compliance in healthcare (β = 0.39, p < .001). Furthermore, international standards such as ISO/IEC 27001 and the NIST Cybersecurity Framework moderated these relationships, amplifying the positive effects of AI adoption on audit effectiveness (β = 0.55, p < .001). These results highlight that AI and ML are not supplementary tools but central mechanisms that redefine the scope, accuracy, and reliability of cybersecurity auditing. The study contributes theoretically by extending the integration of computational intelligence with governance frameworks and contributes practically by offering evidence-based insights for auditors, regulators, and organizational leaders.