Brazil’s Top Electoral Court Weighs a Bolsonaro Deepfake — and AI’s Limits

Brazil · Politics

Key Facts

Core disputeThe TSE will rule on whether an AI-generated video of Jair Bolsonaro endorsing his son Flávio at the PL convention is illegal, even without intent to deceive.
Video originThe deepfake was shown during the PL party convention in July 2026 and challenged by the PT.
Legal ruleArticle 9-C of the TSE’s propaganda resolution prohibits deepfakes in electoral ads, even if the person portrayed authorized the use.
Decision timingTSE president Kassio Nunes Marques is due to meet ministers on Aug. 13, 2026; a ruling is expected during the week of Aug. 17, 2026.
Public opinionA CNT/MDA poll found 49.8% want all AI-made campaign videos banned, while 34.3% would ban only AI content that spreads fake news — meaning 84.1% support some restriction.
Unresolved questionThe court must decide whether AI political content is automatically irregular or only illegal when it deceives voters or causes harmful electoral effects.

Brazil’s top electoral court is about to decide whether an AI deepfake of Jair Bolsonaro is illegal on its face — and that ruling could reshape every AI-generated campaign ad in Latin America.

You may have seen the video by now — a slick, convincing clip of former president Jair Bolsonaro appearing to endorse his son, Senator Flávio Bolsonaro — the PL’s presidential candidate — at the party convention in July 2026. There’s just one catch: Bolsonaro never said those words. The footage was generated with artificial intelligence, and now Brazil’s Superior Electoral Court (TSE) is set to rule on whether that alone makes it illegal. The TSE deepfake case is not about whether the video lied — it’s about whether AI-generated political content is banned even when it isn’t used to deceive or attack. The decision, expected the week of Aug. 17, 2026, could set a precedent for every election in the region.

The deepfake and the challenge

The video in question was aired during the PL convention in July 2026, showing a synthetic version of Jair Bolsonaro voicing support for Flávio Bolsonaro’s candidacy. The PT-led Federação Brasil da Esperança, the coalition of President Luiz Inácio Lula da Silva, filed a formal challenge with the TSE, arguing the clip violated electoral rules. The defense, however, has pushed back with a striking argument: the lack of authorization from the person portrayed is “absolutely indifferent” to whether the content qualifies as a deepfake. What matters, they say, is whether the video is false or misleading.

That position puts the TSE in a difficult spot. The rule at stake is Article 9-C of the TSE’s propaganda resolution, which prohibits deepfakes in electoral advertising even if the person reproduced in the video authorized the use — when the content is meant to harm or benefit a candidacy. The question is whether that prohibition is automatic or conditional. O Globo reports that the current dispute could force the TSE to define whether these cases are automatically irregular or only illegal when there is intent to deceive voters or produce harmful electoral effects.

The legal question at the heart of the TSE deepfake case

This is not a simple case of “fake news.” The deepfake of Bolsonaro was not designed to smear an opponent or spread a false story about a candidate’s health or criminal record. It was, on its face, a promotional tool — an endorsement that never happened, used to boost a family member’s campaign. The core legal question is whether the mere use of AI to fabricate a political statement is inherently illegal, regardless of intent or effect.

If the TSE rules that the video is automatically irregular, it would mean that any AI-generated content in a campaign — even a clearly fictional parody or a hypothetical scenario — could be grounds for sanctions. If the court instead requires proof of deceptive intent, then the Bolsonaro video might be allowed, opening the door for more sophisticated AI tools in future elections. The distinction matters far beyond this one case. Political parties across Latin America are already experimenting with AI for ads, avatars, and voice cloning. A strict ruling in Brazil would send a signal to courts in Argentina, Mexico, and Colombia that AI-generated political content is presumptively illegal.

The likely split and the stakes

According to reports, TSE president Kassio Nunes Marques planned to meet the other justices on Wednesday, Aug. 12, 2026, with the judgment expected for the week of Aug. 17. Some reports describe a possible decision around Thursday, Aug. 13–14. The court is likely to be divided. Nunes Marques, a Bolsonaro appointee, may lean toward a narrower interpretation that requires proof of harm. Other justices, particularly those aligned with a stricter reading of electoral rules, may argue that the rule is clear: deepfakes are banned, period. The split reflects a broader tension in Brazilian law between protecting free political speech and preventing manipulation.

The stakes go beyond the Bolsonaro family. Brazil’s 2026 election is already saturated with AI-generated content, and the CNT/MDA poll shows the public is wary: 49.8% of respondents said no video or election ad produced with AI should be allowed, while 34.3% favored banning only AI content that spreads fake news about candidates. That means 84.1% supported some form of restriction. The TSE’s ruling will effectively decide whether Brazil adopts a blanket ban or a harm-based standard. For investors and expats watching Latin America’s largest economy, this is not a niche legal debate — it’s a test of whether democratic institutions can keep pace with technology that can fabricate reality at scale.

Why this matters for you in Latin America

If you live in or invest in Latin America, you should care because Brazil’s electoral courts are among the most influential in the region. When the TSE sets a rule, other countries often follow. A strict ban on AI deepfakes in campaigns could reduce the risk of election manipulation, which is good for political stability and, by extension, for business confidence. On the other hand, an overly broad ban could stifle legitimate creative uses of AI in political communication, making campaigns more expensive and less innovative. Either way, the ruling will affect how political parties, media companies, and tech platforms operate in Brazil — and that has ripple effects across the continent.

There is also a practical angle. If you are a digital marketer, a media buyer, or a political consultant working in Latin America, this decision will define what you can and cannot produce. If you are an investor in tech or media companies that serve political campaigns, the regulatory environment in Brazil will directly impact your risk assessments. The TSE deepfake case is not a theoretical exercise; it is a concrete decision that will shape the rules of engagement for the next election cycle and beyond.

Frequently Asked Questions

What exactly is the TSE deepfake case about?

The case involves a campaign video shown at the PL convention in July 2026 that used AI to create a fake version of Jair Bolsonaro endorsing his son Flávio. The PT challenged it, and the TSE must decide whether the video is illegal simply because it is AI-generated, or only if it was intended to deceive voters.

When will the TSE rule on this case?

Reports indicate that TSE president Kassio Nunes Marques planned to meet other justices on Wednesday, Aug. 12, 2026. The judgment was expected to be scheduled for the week of Aug. 17, 2026, with some reports pointing to a possible decision around Thursday, Aug. 13–14.

What does the current TSE rule say about deepfakes?

Article 9-C of the TSE’s propaganda resolution prohibits deepfakes in electoral advertising, even if the person reproduced in the video authorized the use, when the content is meant to harm or benefit a candidacy. The open question is whether this prohibition is automatic or requires proof of deceptive intent.

Connected Coverage

Sources: TSE; Estadão; G1; BBC News Brasil; Valor; Correio Braziliense.

This article was produced by The Rio Times’ automated newsroom system. How we use AI · Report an error