Symbolic interactionism provides a robust framework for understanding how meaning emerges through social exchange—a process now transformed by algorithmic platforms. This study examines how symbols within standout frames of TikTok videos influenced engagement during Taiwan’s 2024 presidential election. Using Vision and Large Language Models, we identified visually salient moments and categorized them based on social, cultural, and political symbolic content. Our findings reveal that videos containing multiple symbol types generated significantly higher engagement metrics and emotional responses. Cultural symbols, especially when combined with political imagery, showed the strongest association with trust-building behaviors. This research extends symbolic interactionism theory to digital environments while advancing methodological approaches for analyzing visual political communication in short-form video. The results demonstrate how TikTok’s algorithmic amplification of symbolically rich content shapes political participation during democratic processes.