An AI-generated version of Jair Bolsonaro appeared at his son’s presidential campaign launch with a clear disclosure that it was synthetic, but the video has still triggered a legal challenge over how political deepfakes should be treated during elections.
WHAT’S HAPPENING
Former Brazilian President Jair Bolsonaro appeared virtually at the campaign launch of his son, presidential candidate Flávio Bolsonaro, through an AI-generated recreation of his image and voice.
The synthetic Bolsonaro explicitly told viewers that the appearance was not real and that artificial intelligence had been used to recreate him.
That disclosure is important because Brazil requires political content created or manipulated with AI to be clearly identified.
But opposition parties have challenged the video before Brazil’s electoral authorities, arguing that the synthetic appearance may still violate election rules or allow Jair Bolsonaro to circumvent restrictions placed on his public communications.
The legality of the video has not yet been definitively resolved.
WHY IT MATTERS
Political deepfake regulation has largely focused on deception.
This case presents a more complicated problem:
What happens when everyone knows the politician is fake?
The Bolsonaro avatar was labeled as AI-generated, yet it still allowed the likeness, voice and political influence of a major political figure to appear at a campaign event without that person physically participating.
That means disclosure alone may not resolve every election-integrity concern.
AI could allow campaigns to recreate candidates, political allies or public figures at enormous scale while maintaining that voters were never technically deceived because the content carried an AI label.
WHO BENEFITS
Political campaigns gain powerful new ways to deploy recognizable figures without requiring them to appear physically.
AI media companies could see increasing demand for synthetic video, voice and avatar technology during elections.
Regulators and election authorities gain an important real-world test case for determining where legitimate synthetic media ends and unlawful political manipulation begins.
Voters benefit when AI-generated political content is clearly disclosed rather than presented as authentic footage.
WHO LOSES
Candidates and public figures could lose control over how their image and voice are used if synthetic replicas become common political tools.
Election regulators face increasingly difficult enforcement questions that existing laws were not originally designed to answer.
Voters could also face a flood of synthetic political material where disclosure is technically present but the emotional or persuasive effect remains powerful.
Campaigns that violate emerging AI rules could face serious legal and electoral consequences.
WHAT HAPPENS NEXT
Brazil’s electoral authorities will now help determine whether disclosure is enough—or whether certain uses of synthetic political figures can still cross the line.
Brazil already requires explicit identification of AI-generated political material and has added additional restrictions on synthetic content surrounding elections. (TSE)
The European Union is moving in a similar direction. As of August 2, 2026, Article 50 of the EU AI Act requires disclosure when deepfake image, audio or video content has been artificially generated or manipulated. (Eur-Lex)
The United States currently has federal campaign-ad disclaimer requirements identifying who paid for certain political communications, but those rules are not the same as a nationwide requirement specifically labeling AI-generated political deepfakes. (FEC.gov)
The Bolsonaro case exposes the next problem for election law:
The hardest political deepfakes to regulate may not be the ones pretending to be real—they may be the ones openly admitting they are fake while still carrying the political power of the person they replicate.
