Muhammad Ali Khan

Mr. Muhammad Ali Khan

orcid
OT/ICS Cybersecurity

Research Focus & Metrics

Academic Career Length

1Year (2026-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
15 May 2026 in SSRN Electronic Journal
Verified
Preprint
Open access
Muhammad Ali KhanMuhammad Ali Khan
Muhammad Ali Khan
Artificial Intelligence (AI) is rapidly becoming foundational to cybersecurity operations across Industrial Control Systems (ICS) and Operational Technology (OT) environments, enabling anomaly detection, automated monitoring, and accelerated incident response. However, as industrial defenders increasingly rely on AI-driven security models, a new and largely underestimated threat has emerged: adversarial AI in ICS/OT—the deliberate manipulation, evasion, or poisoning of machine learning systems to create silent detection failures while physical process compromise unfolds undetected. This paper examines how adversarial AI techniques uniquely manifest in industrial environments, where deterministic protocols, stable process baselines, and tightly coupled cyber-physical operations create attack surfaces fundamentally different from enterprise IT systems. It analyzes three primary adversarial vectors relevant to industrial settings: data poisoning of operational telemetry, model evasion through protocol-conforming manipulation, and industrial signal manipulation designed to preserve apparent operational normalcy while degrading physical process integrity. The paper further argues that conventional IT-native AI security models are structurally misaligned with OT requirements because they fail to incorporate industrial protocol awareness, process-physics context, adversarial robustness testing, and explainable detection logic. To address this gap, this work proposes a five-layer practitioner defense framework consisting of: (1) protocol-specific behavioral baselining, (2) AI pipeline integrity validation, (3) cross-verification through deterministic rule-based detection, (4) adversarial testing of OT AI models, and (5) alignment with established industrial security frameworks including NIST SP 800-82 and ISA/IEC 62443. It also maps adversarial AI techniques to the MITRE ATT&CK for ICS framework, positioning AI-targeted manipulation as an extension of established nation-state and advanced persistent threat tradecraft into machine-learning-enabled operational environments. The paper concludes that in critical infrastructure sectors—including energy, manufacturing, water, transportation, and telecommunications—the future of cyber defense is increasingly machine versus machine, where AI-driven defense systems themselves become targets. Organizations deploying AI in ICS/OT without adversarial hardening may create a false sense of security, leaving operational environments vulnerable to stealthy compromise at physical-process scale. This research provides a practical framework for strengthening resilience against emerging adversarial AI threats to critical infrastructure.

23 July 2026 in SSRN Electronic Journal
Verified
Preprint
Open access
Muhammad Ali KhanMuhammad Ali Khan
Muhammad Ali Khan
Artificial intelligence (AI) is increasingly being deployed across operational technology (OT) and industrial control systems (ICS) to support predictive maintenance, anomaly detection, process optimization, and autonomous control. While these capabilities improve efficiency, they also introduce risks that current cybersecurity, safety, and compliance assessments evaluate separately, resulting in inconsistent risk ratings and fragmented governance. Although frameworks such as the NIST AI Risk Management Framework, ISO/IEC 42001, ISO/IEC 23894, ISO 31000, IEC 62443, the NIST Cybersecurity Framework 2.0, and NIST SP 800-82r3 provide valuable guidance, none offers a unified approach to assessing AI risk in critical infrastructure. This paper presents AIRA-CI, an eight-phase AI Risk Assessment Framework for Critical Infrastructure developed using the Design Science Research methodology. The framework introduces the Residual AI Risk Score (RARS), which combines consequence-driven scoring, an AI Risk Amplifier, and a non-compensatory safety gate to better evaluate AI-enabled OT systems. AIRA-CI is validated through standards mapping, comparative analysis, and a water treatment case study, demonstrating a practical, standards-aligned approach that improves the consistency, transparency, and resilience of AI risk assessment in critical infrastructure.

29 July 2026 in SSRN Electronic Journal
Verified
Preprint
Open access
Muhammad Ali KhanMuhammad Ali Khan
Muhammad Ali Khan
Industrial control systems (ICS) and Operational Technology (OT) are up against fast-evolving threats that are advancing faster than the security measures designed to protect them. Hackers are now using the same features that make industrial systems reliable, that is, predictable behaviour, legacy systems, and the need to stay running, avoiding being detected and staying hidden. Artificial Intelligence (AI) can detect and prioritise threats in OT environments efficiently. However, using AI effectively requires plans to be made and executed very carefully due to AI having unique challenges of its own. This article introduces the S.H.I.E.L.D Framework: A Six-layer cybersecurity model for OT/ICS that combines AI-based threat detection, robust protection, zero-trust principles, risk prediction, and industry-standards compliance.

29 July 2026 in SSRN Electronic Journal
Verified
Preprint
Open access
Muhammad Ali KhanMuhammad Ali Khan
Muhammad Ali Khan
The prevailing narrative in artificial intelligence positions creativity and content generation as the primary value drivers of AI systems. However, in cybersecurity, this perspective fundamentally misunderstands where transformative power actually resides. The genuine competitive advantage in enterprise security comes not from what AI generates, but from what it predicts. Modern AI systems, at their core, are probabilistic prediction machines that excel when trained to anticipate threats before they materialize. This paper examines how predictive AI systems transform cybersecurity by enabling organizations to predict system failures, user behavior, fraudulent transactions, insider threats, and emerging zero-day exploits. We explore the critical tension between predictive accuracy and explainability, demonstrating that trustworthy prediction where organizations understand the reasoning behind AI predictions creates the conditions for AI-powered security to scale throughout enterprises. The article synthesizes real-world examples from leading organizations, provides practical implementation frameworks, and addresses the ethical dimensions of deploying predictive security systems. Ultimately, civilizations have always rewarded prediction, and as AI compresses predictive capability into software, the future belongs to organizations that master prediction with trustworthiness.

29 July 2026 in SSRN Electronic Journal
Verified
Preprint
Open access
Muhammad Ali KhanMuhammad Ali Khan
Muhammad Ali Khan
Artificial intelligence is undergoing a fundamental transition from information generation to behavioral forecasting. While public discourse remains focused on generative capabilities, predictive systems are increasingly being designed to anticipate human actions, forecast behavioral patterns, and model psychological vulnerabilities at scale. This paper examines the emergence of behavioral forecasting as a dominant paradigm in AI development and analyzes its implications across cybersecurity, operational technology, critical infrastructure protection, and national security. Through systematic analysis of current applications and emerging threat vectors, this research demonstrates that behavioral forecasting represents one of the most consequential technological developments of the twenty-first century. The ability to predict human behavior with increasing accuracy fundamentally reshapes the attack surface in cybersecurity, extends the scope of infrastructure vulnerability beyond traditional technical domains, and introduces unprecedented risks for social engineering, election interference, and cognitive warfare. Simultaneously, predictive systems offer substantial defensive capabilities in anomaly detection, insider threat identification, and infrastructure resilience. The paper identifies critical challenges including explainability of predictive models, ethical governance frameworks, and the need for transparency mechanisms in high-stakes applications. Evidence suggests that organizations capable of developing robust behavioral forecasting systems will gain substantial strategic advantages, while those failing to address associated risks face heightened exposure to sophisticated, algorithmically-optimized attacks. The future of cybersecurity increasingly depends not on reactive defense but on predictive intelligence.