Brazil Defines AI Deepfakes Just Weeks Before Election Amid Surge in Synthetic Campaign Content

Brazil's Electoral Court Tackles AI Deepfakes Five Weeks Before Vote

Brazil's top electoral court, the Tribunal Superior Eleitoral (TSE), issued a landmark ruling Tuesday defining what constitutes AI deepfakes in political advertising, just weeks before the country's October 4 election1

1

2

2

. The decision comes as Brazilian authorities grapple with how to keep AI from undermining the hotly contested presidential race between President Luiz Inácio Lula da Silva and opposition candidate Senator Flávio Bolsonaro1

1

. In a 5-2 vote, Brazil's electoral court defined deepfakes as "synthetic content produced or manipulated via artificial intelligence or equivalent technology" with sufficient realism "to create, reproduce or alter the image, voice or speech of any living, deceased or fictional person"2

2

4

4

. This new standard for AI deepfakes aims to curb misinformation and voter manipulation as AI-generated content floods social media feeds across the country.

The Bolsonaro Avatar Case That Sparked Regulatory Action

The ruling emerged from a controversial case involving Flávio Bolsonaro, who featured an AI avatar of his father, former President Jair Bolsonaro, at his Liberal Party convention in São Paulo in July1

1

2

2

. Jair Bolsonaro, sentenced to 27 years in prison for leading a coup attempt and currently under house arrest, is prohibited from addressing the public2

2

4

4

. The AI-generated version appeared to urge attendees "to open arms" for his son's presidential bid2

2

. While the electoral court voted 4-3 not to punish Senator Bolsonaro, ruling that the deepfake did not constitute a political ad for public airing but rather content for a private event before his candidacy was officially announced, the case highlighted a critical gap in AI use in campaigns2

2

4

4

. The avatar explicitly identified itself as a simulation, sitting awkwardly between disclosure and deception and challenging existing frameworks designed primarily to target fabricated content that misleads voters3

3

.

AI Dominates Digital Campaigns Despite Regulatory Efforts

AI has already dominated much of the digital campaign, with politicians aligned with the leading candidates sharing hundreds of social media posts formally labeled as AI-generated over a 15-day period, according to Instituto Democracia em Xeque (Democracy in Check Institute)1

1

5

5

. While Lula's campaign largely shuns AI-generated content, Bolsonaro's camp has embraced it enthusiastically, sharing AI-produced campaign songs, videos of candidates as dancing dolls, and even an AI animation of Bolsonaro as a chiseled figure that sparked numerous memes1

1

5

5

. The TSE already received 19 cases involving AI use in the first 11 days of this year's campaign1

1

5

5

. In early August, TSE Justice André Mendonça ordered removal of a video portraying Bolsonaro alongside Daniel Vorcaro, a former banker at the center of a massive fraud scandal, in situations that never occurred1

1

5

5

.

Comprehensive Brazil AI Regulation Framework Takes Shape

Brazil AI regulation now includes mandatory labels on AI-generated content, bans on deepfakes and candidate recommendation systems, and a blackout on publication or paid promotion of new AI-generated material immediately before and after voting1

1

5

5

. These expanded restrictions aim to regulate AI ahead of election day and protect the democratic process1

1

. The rules will apply to vast amounts of content, with broadcasters required to set aside 14 minutes daily for nearly 1,000 30-second ads, plus 12 half-minute blocks reserved exclusively for presidential candidates2

2

4

4

. "It is not the tool itself that is the problem. It is its potential use in certain contexts that may lead to dangerous outcomes for our democracy," said João Victor Archegas, legal coordinator at the Institute of Technology and Society (ITS) of Rio de Janeiro1

1

5

5

.

Chatbots and Generative AI Systems Pose New Challenges

Another area of concern involves chatbots swaying voters, with a survey by Projeto Brief finding that nearly 63% of voters were willing to consult AI tools about candidates1

1

5

5

. The TSE's rules for 2026 include a ban on generative AI systems, such as ChatGPT, from recommending or ranking candidates, even when users explicitly request it1

1

5

5

. However, a study by ITS Rio found that seven of the most widely used chatbots in Brazil ranked or ordered candidates in 90% of their responses, highlighting significant enforcement challenges1

1

5

5

. With AI already part of the daily lives of 75% of Brazilians according to a Fundação Itaú Observatory survey, the technology influences voters through fake testimonials about candidates and fabricated polls1

1

5

5

.

Global Implications and Enforcement Questions

The timing presents difficulties, as campaigns, platforms and regulators have just five weeks to adjust to the framework Brazil's electoral court has established3

3

. Enforcement will prove challenging, particularly when content moves through private messaging groups rather than public social networks3

3

. European regulators have practical reasons to monitor Brazil's approach, as the EU's own transparency requirements rely on similar assumptions about disclosure3

3

. The blackout provision on synthetic content around election day represents a significant step beyond existing European rules, which generally focus on transparency and circumstances where synthetic content can mislead3

3

. The Bolsonaro case demonstrates that content can be completely transparent about its synthetic nature yet remain politically persuasive, putting pressure on disclosure-based frameworks being developed under the EU's AI Act3

3

. Watch for how effectively Brazil enforces these new standards for AI deepfakes and whether other democracies adopt similar approaches to protect electoral integrity in an era where political advertising increasingly relies on AI-generated content.