Abstract:  The article presents an interdisciplinary analytical framework contributing to the growing research field of visual political communication, focusing on the case of the social media images published by Italian politicians during the 2024 European elections campaign (May–June 2024). In the first part, the article outlines the context of the analytical framework at the intersection of three main research fields: political communication, in particular the study of electoral campaigns via social media; visual culture and communication, precisely the analysis of the visual representation, self‐representation, and counter‐representation of political leaders; and computer science, in particular the application of machine learning techniques for computer vision to recognize and categorize visual political content. In the second part, the article offers an application of the analytical framework by sharing some empirical results of a quantitative and qualitative analysis of the visual content published by 21 Italian political actors on Facebook and Instagram during the campaign, focusing on their main visual formats, themes, and strategies of representation of political leadership. In the analysis, deep learning models are also employed to detect specific image characteristics by cross‐referencing their outputs with manual cataloguing performed on the same images and for the same attributes. In the end, on the basis of the research carried out, the article suggests possible paths for future interdisciplinary analysis of online visual political communication.

Keywords:  AI images; computer vision; digital campaigns; electoral campaigns; political communication; political leadership; visual political communication

Published:   23 September 2025

Issue:   Vol 13 (2025): Electoral Communication: European Elections in Times of (Poly)Crises (In Progress)