A study identified 18 AI-generated political influencer avatars used to comment on politics during Brazil's 2024 municipal elections [1], [2].
This development highlights a critical gap in electoral regulations as synthetic personas are used to mimic human interaction. Because these avatars can operate at scale without traditional accountability, they pose a significant risk to the integrity of democratic discourse.
The research was conducted by the Observatório Inteligência Artificial nas Eleições, which found these characters active across social media platforms including TikTok, Instagram, and X [1], [2]. According to the study, political actors and campaign teams deployed these avatars to amplify specific messages and shape public opinion [3], [4].
These digital personas are designed to look and act like real people to deceive the electorate. “These avatars are programmed to reproduce political speeches, confusing voters who believe they are interacting with real people,” said Dr. Mariana Oliveira, a researcher with the Observatório Inteligência Artificial nas Eleições [4].
The emergence of these synthetic influencers suggests that traditional deepfake regulations may be insufficient. While many rules focus on altered videos of real candidates, these avatars are entirely fabricated entities created to build trust with audiences through simulated authenticity.
Legal experts and observers suggest that the current framework cannot keep pace with this technology. “The use of influencers created by AI in campaigns demonstrates a new frontier of disinformation that is still not adequately regulated,” said Andréia Sadi, a blogger for G1 [1].
The study indicates that these 18 identified avatars were part of a broader trend of AI integration in the 2024 cycle [1], [2]. Some reports suggest dozens of such accounts were active, though the specific count of 18 represents the verified set from the observatory [1].
““These avatars are programmed to reproduce political speeches, confusing voters who believe they are interacting with real people.””
The shift from deepfaking existing politicians to creating entirely synthetic 'influencers' represents a tactical evolution in digital campaigning. By creating personas that do not exist in the real world, political actors can deploy aggressive or polarizing rhetoric while maintaining plausible deniability. This challenges the current legal definition of 'candidate' and 'campaign spending,' as it becomes harder for regulators to track who is funding the creation and promotion of these artificial voices.
