This paper introduces Semantic Betweenness Centrality (SBC), a novel family of centrality measures for analyzing complex semantic networks. Unlike classical betweenness centrality, which relies solely on topological shortest paths, SBC evaluates the importance of nodes based on their participation in semantically optimal pathsthose that are most resilient, scalable, resource-efficient, or rapid in realization. We demonstrate that node centrality is not absolute but contextdependent: a node critical for resilient influence may be insignificant for fast or low-cost strategies. As a case study, we construct a comprehensive semantic network of cognitive warfare comprising 290 nodes and 310 directed links, modeling pathways from strategic objectives to psychological consequences. The network is built using a "swarm of virtual experts" approach, leveraging Large Language Models (LLMs) to generate diverse, interdisciplinary insights through role-based prompting. While LLMs facilitate rapid, scalable, and semantically rich network expansion, the proposed SBC framework remains methodologically independent and applicable to any domain with semantically annotated paths. Results show that redefining optimality by semantic criteria leads to dramatic shifts in node rankings, underscoring the limitations of traditional topological metrics. The SBC approach enables goal-oriented analysis of influence structures, with applications in information warfare, strategic communication, and cognitive security. This work advances network analysis by integrating semantic reasoning into structural metrics, offering a multidimensional perspective on influence and control in complex systems. Keywords: Semantic Centrality, SBC, LLM, Betweenness Centrality, Cognitive Warfare, Information Chains, Semantic Network