Brazil has 18 political influencers created by AI on social media
A study by the AI Observatory in Elections revealed the advancement of the use of characters created by artificial intelligence on social networks to influence political debate - a type of "fictitious influencers".
According to the research, these characters appear as supposed voters, influencers, political commentators, presenters and even popular leaders. They are used to comment on political topics, criticize authorities and disseminate content about the national scene.
The survey identified 18 avatars used for this purpose active on social networks. Of this total, 11 profiles do not show any indication that the content was produced using AI, equivalent to 61% of the pages analyzed.
According to the observatory, in the 7 cases in which there was some type of signaling, the warnings appeared through markers on the platforms themselves, watermarks from the tools used or through hashtags inserted in the publications.
This type of content circulates mainly on TikTok and Instagram, with 6 cases on each platform. YouTube appears next, with 3 cases. Content was also identified on X, Kwai and Facebook, with 1 case on each network.
The study also points out that 78% of all material disseminated by these characters contains false or distorted allegations involving politicians, democratic institutions and public policies.
Among the true content, the survey identified that the majority has a tone of criticism or political attack. In total, 13 cases were used to criticize politicians, governments, institutions or public policies.
In an interview with CNN Brasil, the director of Aláfia Lab - one of the companies responsible for the observatory - Rodrigo Carreiro, stated that the increased presence of fictional characters in the political debate favors hate speech.
"This is a discourse of disinformation that can be massively shared, mainly because people are consuming this type of content more and more. We are experiencing an artificial intelligence fad."
The "Dona Maria" case
Among the cases that attracted the most attention is that of the fictional influencer known as "Dona Maria". The character is portrayed as a black woman and was created artificially to publish content critical of the government of President Luiz Inácio Lula da Silva.
The profile was first identified by the observatory in August 2025 and, since then, it has published more than 400 videos on social media.
In addition to attacking the federal government, the character's content also targets ministers of the Federal Supreme Court (STF), such as Alexandre de Moraes, Cármen Lúcia and Luís Roberto Barroso.
Publications include criticisms of the functioning of democratic institutions and questions about government public policies, including Bolsa Família.
Parties' reaction
According to CNN Brasil, the PT (Workers' Party), the PV (Green Party) and the PCdoB (Communist Party of Brazil) sued the TSE (Superior Electoral Court) to take down these profiles created by artificial intelligence and which are used to criticize the Lula government.
According to the representation, "it is a clear tool of political propaganda, used consciously and deliberately by a certain unknown person to, under anonymity, publish untruths, decontextualizations, commit crimes against honor, while showing praise towards certain political figures from another political spectrum."
Source: CNN