Viktoriia Shcherba
, Mykhailo Akimov
, Inna Vartyletska
, Olha Kryshevych
, Mariia Diakur
Generative AI has accelerated deepfake crime beyond the limits of traditional victim-centered criminal qualification. This study develops a dual-scale framework that links personal and systemic harm to criminal-judicial decision-making. Using classification, topology building, cross-impact analysis, judicial content analysis, regression, scenario modeling, and UML formalization, the study shows that automation, diffusion, and institutional targeting drive the most severe outcomes.