Develop an empirical model to examine the phenomenon of Cross-Border Electoral Interference triggered by US-party-centred media influence operation activities. Based on a public-domain collection of 62 observed cases occurring from 2010 through 2024; coding six influence channels, five electoral narrative targets, three election stages, five regional environments, and three tiers of national security significance. Analysis is organised through the National Security-Constrained Media Influence Graph (NS-CMG) to link actors’ loci, media channels, narrative targets, election periods, hosts-States of vulnerability, attributive ambiguities, and evidence convergences. Construct two indexes, namely the influence exposure index of channel-narrative intensity and the cross-border risk index of review salience. The results show that platform advertising and data-targeting have risen from 19.8 to 87.6 on a 0-100 exposure scale from 2010 to 2024; Allied media narrative syndicated information has increased from 20.4 to 84.1. The strongest contribution appears in election-integrity response linked to election-integrity claims, with a mean value of 36. The three-dimensional risk surface shows that, at this point of the combination index of high saliency + attribution ambiguity ≥ 4.17; At the event level, F1 is 0.82, ROC-AUC is 0.90, and the expected calibration error is -0.048. The subsequent forms add restrictions on reviewing the external effects of countries/regions separately in terms of descriptions, attributions, and reasonsableness.
