Brazil's October 2026 election will be the first major national vote in the country where generative AI is extensively used to create synthetic social media personas, challenging regulators and electoral authorities to define and control AI-driven misinformation at scale.
- Generative AI persona 'Dona Maria' reaches over 100 million people pre-election.
- Brazil’s Senate passed an AI regulatory bill influenced by the EU AI Act, still pending lower house approval.
- Electoral Court rules on AI in elections remain vague, complicating enforcement against synthetic content.
What happened
Ahead of Brazil's October 2026 presidential election, a viral AI-generated social media persona known as Dona Maria has attracted massive attention by commenting on political and economic issues. Created by an Uber driver using tools like Google Gemini and Flow, Dona Maria mimics an elderly Black woman, a key voter demographic expected to influence the election outcome. Despite the account’s fictional nature, its videos have garnered millions of views and sparked widespread belief that the persona is real.
This extensive use of generative AI to craft synthetic political commentary is unprecedented in Brazilian elections. The phenomenon underscores how accessible AI technology can be leveraged by ordinary individuals to shape political narratives. However, it also exposes gaps in Brazil’s evolving AI regulatory environment and raises questions over the definition and regulation of AI-created content in the election context.
Why it matters
The case of Dona Maria highlights the difficulty in classifying AI content—whether it qualifies as a deepfake or another form of synthetic media—and regulating it accordingly. Without a clear legal framework, inconsistent enforcement or legal interpretations could ensue, potentially impacting voter perceptions and electoral fairness in one of the world’s largest democracies.
What to watch next
Observers will be closely monitoring how Brazil’s Chamber of Deputies responds to the Senate's AI regulation bill and whether it adopts provisions that concretely address generative AI in political content. Additionally, clarifications from the Superior Electoral Court on what defines prohibited synthetic media, especially concerning personas like Dona Maria that do not impersonate real individuals but simulate societal voices, will be pivotal.
The impact of AI-driven synthetic content on voter behavior and misinformation dynamics during the election campaign may shape future regulatory approaches not only in Brazil but across other democracies confronting the rapid evolution of AI tools. How authorities balance transparency, free speech, and electoral integrity in regulating AI will be critical to watch.