Preprint / Version 2

RUMOR EFFECTS IN THE AGE OF ARTIFICIAL INTELLIGENCE: A TURN FOR THE PEOPLE?

##article.authors##

DOI:

https://doi.org/10.1590/SciELOPreprints.14328

Keywords:

Rumor-effect, Political discourse, Generative Artificial Intelligence, Technical materiality, Political body

Abstract

This article revisits the notion of the rumor effect in order to examine the functioning of political and ordinary discourse within their digitally mediated material imbrication. It analyzes audiovisual productions that employ generative Artificial Intelligence (AI) to simulate and satirize the bodies and voices of “the people” and political representatives. The study is grounded in the hypothesis that the technical capacity to simulate the people’s voice and body intensifies and complexifies the rumor effect by displacing the regimes of evidence and credibility that underpin discursive circulation. The empirical corpus comprises the “Taxação BBB” campaign developed by the Workers’ Party and the satirical pieces disseminated by the profile “Brasil Sátira do Poder,” both of which mobilized AI-generated videos within the context of the same political dispute. Through qualitative discourse analysis, the study proposes a reconceptualization of the rumor effect, understanding it as a discursive operation traversed by the technical materiality of the digital environment, in which image-based simulation actively participates in the constitution of truth effects. The findings indicate that, in these productions, the body of the people metaphorizes the political, while the political body metaphorizes the people, thereby enabling a displacement of their respective enunciative positions. Consequently, the rumor effect emerges not merely as a phenomenon of circulation but as a material-discursive process shaped by the affordances of generative AI.

Downloads

Download data is not yet available.

Posted

12/23/2025 — Updated on 03/02/2026

Versions

How to Cite

RUMOR EFFECTS IN THE AGE OF ARTIFICIAL INTELLIGENCE: A TURN FOR THE PEOPLE?. (2026). In SciELO Preprints. https://doi.org/10.1590/SciELOPreprints.14328 (Original work published 2025)

Section

Linguistic, literature and arts

Plaudit

Version justification

Esta versão atualiza o preprint para adequação às exigências técnicas da revista Cadernos de Linguística, incluindo ajustes formais, inserção de quadros e descrições ALT, atualização das declarações (Declaração atualizada de uso de dados, CRediT, Uso de IA, Conflito de Interesses) e inclusão de DOI para o pacote de dados depositado no Zenodo.

Data statement