This paper proposes a novel approach to analyzing node centrality in information networks, based on the semantic diversity of paths. It is shown that traditional metrics, such as betweenness centrality, are limited by topological logic: they account only for the number of shortest paths, ignoring their semantic content. In social and cognitive networks, paths possess semantic properties, such as resilience to countermeasures, scale of impact, resource intensity, or legitimacy.The paper examines the concept of semantic betweenness centrality (SBC) as a distinct metric for each semantic optimality criterion. To overcome the limitations of the deterministic approach, an extended model is proposed: Probabilistic Semantic Betweenness Centrality (PSBC). This model integrates a logical-probabilistic approach, evaluating centrality proportionally to the probability of realizing a specific semantic path. This allows for filtering out "theoretically optimal" but improbable chains of influence, focusing instead on genuinely threatening or highly effective trajectories.Using the example of a cognitive warfare semantic network, it is demonstrated how changing the analysis objective and accounting for probabilistic characteristics radically alters the hierarchy of key concepts. This indicates the relative nature of centrality and paves the way for the creation of context-dependent and more precise metricsof influence. The results are significant for the analysis of information-psychological operations, the development of mechanisms to counter cognitive warfare, and strategic planning in the field of national security Keywords: semantic-probabilistic centrality, SBC, PSBC, betweenness centrality, cognitive warfare, cybersecurity, information chains, semantic network