How short videos shape natural disaster insurance purchase intentions in the face of natural hazards: a PLS-SEM–ANN analysis

The rapid rise of short-form video platforms such as TikTok, has revolutionized information dissemination, particularly among younger, tech-savvy audiences. This study, the first of its kind, investigates how short videos influence individuals' purchase intentions for natural disaster insurance, with a focus on China. This study proposed a model to examine how short videos, as external stimuli, shape internal states – specifically fear and risk perception – and drive behavioural responses, namely purchase intention. We analyze three types of short video content: natural disaster information (NDISV), disaster prevention information (DPISV), and disaster insurance knowledge (DIKSV). Data from 426 Chinese respondents were analyzed using PLS-SEM and ANN. The results demonstrate that NDISV, DPISV, and DIKSV significantly enhance purchase intentions, with fear acting as a mediator between short video content and risk perception. Risk perception, in turn, positively influences purchase intention, underscoring the role of emotional appeals in risk communication. The ANN analysis revealed nonlinear relationships and identified DIKSV and DPISV as the most influential factors in predicting purchase intention. These findings highlight the potential of short videos as a powerful tool for raising awareness of natural disaster risks and motivating protective behaviours, providing actionable insights for policymakers, insurers, and content creators.