AI Analysis of Visual Metaphors in Political Posters

Posters, as a mode of political communication, play a role in shaping public opinion; its use of visual metaphors helps structure or elaborate on complicated ideological and narrative frameworks. Nonetheless, the common approaches to mine these metaphors are inherently qualitative and labor-intensive, resulting in a limited scalability and lack of objectivity. The present study shows how machine learning and textual analysis can be used in an AI-based framework to detect and classify visual metaphors appearing in ideological political posters across different historical presentational contexts. It's a collection of more than 1,000 posters spanning elections, propaganda and social movements from 1920 to 2025. Using CNNs (Convolutional Neural Networks) and image preprocessing methods in higher accuracy (87 %) were reached for detecting & classifying metaphors and, the method provided better consistency and scope than manual analysis. Startlingly, scholars discover that conservative campaigns are heavily dependent on power-oriented metaphors and imagery, and conservative campaigns mainly rely on humanitarian and environmental metaphors. The study is significant not merely in verifying AI's effectiveness at decoding visual political rhetoric but also with capacity for large-scale, impartial analysis of political communication strategy.