2026 ELECTIONS

Once a trailblazer in electoral AI rules, Brazil is now falling behind

Today’s use of AI imagery is largely about making memes, such as this one picturing Lula and Jair Bolsonaro.

Brazil’s 2024 municipal elections caused alarm bells to ring for authorities and experts around the country, as generative artificial intelligence — especially deepfakes — appeared poised to take electoral information chaos to a new level.

The world has been grappling with a frenzy of social media disinformation for more than a decade now. In Brazil, the 2018 presidential election was marked by the rise of mass-messaging campaigns on WhatsApp. Four years later, in 2022, videos and livestreams spread by influencers and politicians flooded the public discourse with falsehoods.

The next electoral weapon, many feared, would come from digitally manipulated sounds and images, making lies even more convincing. Brazil’s Superior Electoral Court (TSE) rushed to establish rules to curb the practice, moving ahead of many democracies worldwide.

Then led by Supreme Court Justice Alexandre de Moraes, the TSE required electoral advertisements to disclose when they were produced using AI. Deepfakes were banned, with candidates facing disqualification if they were used.

There were serious crimes, with women candidates as the primary targets of sexually explicit deepfakes. But the impact of AI proved less disruptive overall than initially feared. According to a study by the Getulio Vargas Foundation’s Communication School, most AI-generated content that gained traction was satirical and easily identifiable.

The major AI wave, expected in 2024, was thus postponed to 2026. In October, Brazilians will head to the polls to choose a president, state governors, and federal and state lawmakers. However, this time around, Brazilian authorities appear less attentive to the threats of AI than the moment demands.

“We are concerned that the TSE has reached 2026 without addressing the evolution of AI, which is no longer easily detectable by users,” said Carla Rodrigues, coordinator for digital platforms and markets at Data Privacy Brazil. Experts have pointed out that new AI developments will allow fraudsters to rapidly produce phony but close-to-perfect pictures, video or audio of candidates and officials doing or saying almost anything with minimal time…

🔒 This was a free preview; the rest is behind our paywall

Don’t miss out! Upgrade to unlock full access. The process takes only seconds with Apple Pay or Stripe. Become a member.

Why you should subscribe

We’re here for readers who want to truly understand Brazil and Latin America — a region too often ignored or misrepresented by the international media.

Since 2017, our reporting has been powered by paid subscribers. They’re the reason we can keep a full-time team of journalists across Brazil and Argentina, delivering sharp, independent coverage every day.

If you value our work, subscribing is the best way to keep it going — and growing.

Reply

or to participate