Adversarial threats to digital twin technology: a taxonomy of vulnerabilities and attack surfaces

Digital twin (DT) technology is an emerging pillar of Industry 4.0, offering real-time simulation capabilities to forecast, optimize, and improve multiple real-world systems across diverse domains, including healthcare, manufacturing, and smart cities. The deployment of a DT relies on the seamless integration of advanced technologies such as cyber-physical systems, the Industrial Internet of Things (IIoT), edge computing, virtualization infrastructures, artificial intelligence, and big data. While this integration enables unprecedented efficiency and innovation, it also introduces complex security threats that demand significant attention. Among these threats, adversarial attacks pose a particularly insidious risk by exploiting vulnerabilities in machine learning models, IoT systems (Imran and Anjum, Comput Mater Continua 68(2), 2021), and data flows to compromise the accuracy and reliability of digital twins. This chapter probes into the current state of the DT paradigm with a specific focus on adversarial attacks, classifying the potential threats across its functionality layers and operational requirements. We provide a systematic taxonomy of adversarial threats, considering their impact on real-time data processing, simulation accuracy, and system integrity.