As the world experiences its fourth industrial revolution, commonly known as Industry 4.0 (I4.0), practitioners are beginning to engage with I4.0 technologies, such as artificial intelligence (AI) and digital twins (DT). Such technologies are automating and digitizing previously analogue processes, enabling modern and sustainable asset management. Numerous studies have demonstrated the efficiency of those systems from monitoring to decision-making. However, as reliance on I4.0 technologies increases, a major concern arises: are these systems truly trustworthy? On one hand, there are worries about the system's accuracy and uncertainty, and on the other, the threat of malicious attacks from humans. For structural health monitoring (SHM), artificial disturbances can be introduced into an AI-based SHM system, perturbing healthy signals appear damaged, or vice versa, disguising damaged signals as healthy to obscure the true condition of a structure. Such malicious alteration could lead to catastrophic consequences. This paper discusses these by presenting a case study on traffic event-based deep-learning bridge SHMs, presenting methods for attacks and defense. The goal is to attract public attention to the vulnerability of AI-driven asset management systems and the development of defense means.