Memes and AI in Political Radicalisation in Latin America

This presentation examines the memetic communication of the main right-wing populist
movements in different Latin American countries—Brazil, Guatemala, Argentina, and El Salvador
—during electoral competition for their countries’ presidency. The study seeks to understand how
new Generative AI tools (such as Stable Diffusion, MidJourney, DALL-E, GAN or Generated
Faces) affect the dynamics in which these memes are produced and distributed, intensifying
political polarisation and micro-segmented information bubbles.
This study is based on both online and offline research in order to understand how memetic
communication mobilise elements that promote political polarisation and incite violence. Between
September 2022 and February 2024, in the weeks leading up to and following each presidential
election, I collected and analysed visual data employing open-source software. I also conducted
ethnographic fieldwork and digital ethnography during the weeks preceding the elections to capture
online and offline discourses and the affective milieu of each electoral campaign as a means to
contextualise the social media data and the impact of the memetic communication.
Overall, the research shows the central role that Generative AI is playing in the way memetic
communication operates, ultimately reinforcing political polarisation and the legitimisation of
violence against political opponents and other social groups, especially women, LGTBIQ+ and/or
racialised people, who are stereotyped, denigrated and deshumanised.