EXPLAINABLE AI FOR REAL-TIME CYBER THREAT DETECTION IN PAKISTAN’S CRITICAL DIGITAL INFRASTRUCTURE: BALANCING ACCURACY, TRANSPARENCY, AND TRUST

Abstract

The increasing frequency and sophistication of cyber threats targeting critical digital infrastructure necessitate the development of advanced, intelligent, and trustworthy cybersecurity solutions. While Artificial Intelligence (AI)-based intrusion detection systems have demonstrated high effectiveness in real-time cyber threat detection, their lack of interpretability limits operational trust and decision-making reliability. This study investigates the role of Explainable Artificial Intelligence (XAI) in enhancing real-time cyber threat detection within Pakistan’s critical digital infrastructure by balancing accuracy, transparency, and trust. A quantitative explanatory research design was employed, with data collected from 320 cybersecurity professionals working in key infrastructure sectors, including banking, telecommunications, energy, and government IT systems. The results revealed that XAI significantly improves system interpretability, enhances trust in AI-driven security systems, and strengthens decision-making efficiency in Security Operation Centers (SOCs). Moreover, strong positive relationships were found between XAI interpretability, detection accuracy, and operational trust. The study concludes that integrating XAI into AI-based cybersecurity frameworks is essential for developing reliable and transparent intrusion detection systems in high-risk digital environments. The findings provide important implications for cybersecurity practitioners, policymakers, and system developers aiming to strengthen national cyber resilience through explainable and trustworthy AI systems.