The adoption of generative artificial intelligence (AI) has attracted growing attention across industries due to its potential to transform organizational processes and value creation. Despite its high applicability, however, the diffusion of generative AI in the telecommunications industry remains limited. Existing studies have largely focused on identifying individual barriers to AI adoption, providing insufficient understanding of how these barriers interact and form a complex hierarchy of constraints. Addressing this gap, this study investigates the structural interrelationships among barriers to generative AI adoption in the telecommunications industry. Based on a comprehensive literature review and expert validation, fifteen key barriers were identified. Using a Delphi-based Interpretive Structural Modeling (ISM) approach, this study examined the hierarchical influence structure among the barriers. Subsequently, the Matrix Impact Cross-reference Multiplication Applied to Classification (MICMAC) technique was employed to classify the barriers according to their driving power and dependence. The results reveal a four-level hierarchical structure in which environmental barriers play a foundational role. In particular, the absence of alignment in institutional frameworks and technical standards emerges as a root-level barrier exerting strong influence on higher-level constraints. Regulatory uncertainty and concerns about job displacement function as independent drivers linking foundational environmental conditions to execution- level constraints. Most technical, organizational, and economic barriers are concentrated at the intermediate level, forming a highly interdependent execution layer. At the top level, delays and uncertainties in decision-making regarding generative AI adoption appear as outcome-oriented barriers resulting from the cumulative effects of lower-level constraints. By highlighting that barriers to generative AI adoption in the telecommunications industry operate as a structurally connected system rather than isolated factors, this study extends existing adoption research through a structural perspective. The findings provide practical insights for telecommunications firms in prioritizing adoption strategies and offer implications for addressing institutional and regulatory conditions that shape the diffusion of generative AI.

Unlike other facilities, maintaining processes is essential in industrial facilities. Pipe racks, which support pipes of various diameters, are important structures used in industrial facilities. Since the transport process of pipes directly affects the operation of industrial facilities, a fragility curve should be derived based on considering not only the pipe racks' structural safety but also the pipes' transport process. There are several studies where the fragility curves have been determined based on the structural behavior of pipe racks. However, few studies consider the damage criteria of pipes to ensure the transportation process, such as local buckling and tensile failure with surface defects. In this study, an analysis model of a typical straight pipe rack used in domestic industrial facilities is constructed, and incremental dynamic analysis using nonlinear response history analysis is performed to estimate the parameters of the fragility curve by the maximum likelihood estimation. In addition, the pipe rack's structural behavior and the pipe's damage criteria are considered the limit state for the fragility curve. The limit states considered in this paper to evaluate fragility curves are more reasonable to ensure the transportation process of the pipe systems.