Cyber Security Education by integrating Digital Twins and Generative AI[v1] | Preprints.org

Digital Twins (DTs) create virtual replicas of physical systems such as programmable logic controllers (PLCs) and Internet of Things (IoT) devices, enabling real-time monitoring, simulation, and analysis of cyber environments without disrupting operational systems . In cybersecurity education, DTs provide immersive, risk-free environments that replicate IT and operational technology (OT) infrastructures, allowing learners to engage with realistic cyberattack scenarios like phishing, ransomware, and zero-day exploits . This capability supports experiential learning by bridging the gap between theoretical knowledge and practical application .

Generative Artificial Intelligence (GenAI), particularly through Large Language Models (LLMs), enhances these DT-based training platforms by automating tasks such as penetration testing, threat modeling, and incident response . Tools like PentestGPT and PentestAgent leverage LLMs to generate attack vectors, interpret system logs, and suggest remediation strategies in real time . In educational settings, GenAI tailors the complexity of simulated attacks to the learner's skill level and provides contextual, natural language feedback during exercises . The integration of GenAI also supports red and blue team exercises by generating adaptive threat narratives and facilitating role-based attack and defense simulations .

A key implementation of this integration is the Red Team Knife (RTK), a custom penetration testing toolkit aligned with the Cyber Kill Chain model, which guides learners through phases such as reconnaissance and exploitation within a DT-powered environment . The combination of DTs and LLMs enables intelligent, real-time feedback and improves situational awareness by translating complex technical data into understandable insights, benefiting both experts and non-experts . This synergy enhances detection, response, and prevention capabilities by enabling dynamic simulations, automated vulnerability assessments, and adaptive learning pathways .

The framework is currently being piloted in academic environments to develop hands-on skills in vulnerability assessment and security operations, with initial findings indicating a significant improvement in training effectiveness . Beyond education, this approach has potential applications in active threat modeling, diagnostics, and adaptive defense strategies . By integrating human-AI collaboration and addressing socio-technical aspects like social engineering, the model supports a broader concept of Cyber Social Security, where psychological and social vulnerabilities are also considered in training

This research proposes a novel educational framework designed to bridge the gap in Operational Technology (OT) security training by integrating Digital Twins (DT) with Generative AI (GenAI). Recognizing the inherent risks associated with live-fire testing on physical Programmable Logic Controllers (PLCs) and IoT infrastructure, the authors utilize DTs to create high-fidelity, safe-to-fail virtual replicas of industrial environments. These digital environments serve as isolated sandboxes where students and trainees can interact with realistic control systems without the threat of causing physical damage, operational downtime, or safety hazards to critical infrastructure.

The key contribution of this work lies in the augmentation of these digital environments with GenAI capabilities to automate and enhance complex security workflows. By leveraging GenAI, the framework facilitates automated penetration testing and dynamic threat modeling, effectively simulating adversarial behavior with high sophistication. This integration allows for the on-the-fly generation of diverse attack vectors and realistic scenarios, transforming static simulation exercises into dynamic, responsive learning experiences that can adapt to the user's skill level and expose emergent vulnerabilities.

This material matters significantly because it addresses the critical shortage of skilled cybersecurity professionals capable of defending industrial environments against modern threats. By removing the barrier of physical risk and utilizing AI to accelerate the generation of complex attack scenarios, the approach democratizes access to high-quality OT security education. It represents a paradigm shift from traditional, textbook-based learning to an immersive, AI-driven pedagogical model, better equipping the workforce to handle the evolving threat landscape in industrial IoT and critical infrastructure sectors.

Summary: Cyber Security Education via Digital Twins and Generative AI

This research paper explores the integration of Digital Twins (DTs) and Generative AI (GenAI) to enhance cybersecurity education by providing realistic, risk-free environments for security testing. Digital Twins replicate industrial control systems (PLCs) and IoT devices, enabling students and practitioners to experiment with penetration testing, threat modeling, and defensive strategies without disrupting real-world operations. The paper highlights how GenAI can automate the generation of synthetic attack scenarios, adaptive threat intelligence, and personalized learning pathways, making cybersecurity training more dynamic and scalable.

The key contributions include a framework for DT-based security labs that can simulate complex cyber-physical systems (CPS) and an AI-driven approach to threat simulation that reduces reliance on static datasets. By leveraging GenAI, the system can generate realistic attack vectors and countermeasures, improving hands-on learning outcomes. This approach is particularly valuable for bridging the skills gap in cybersecurity, where practical experience is often limited by ethical and operational constraints. The paper underscores the potential of this methodology to revolutionize cybersecurity education by making it more interactive, data-driven, and aligned with real-world threats.

Why It Matters: As cyber threats grow in sophistication, traditional training methods—relying on outdated labs or theoretical knowledge—are insufficient. This research offers a scalable, AI-augmented solution that combines the fidelity of Digital Twins with the adaptability of GenAI, making advanced cybersecurity training accessible to a broader audience. For educators, researchers, and industry professionals, this work provides a blueprint for modernizing cybersecurity curricula to better prepare the next generation of security experts.

Source: [Preprints.org](https://www.preprints.org/manuscript/202511.1203)