AI-generated videos flood Brazil's presidential election campaign

AI-generated videos ranging from humorous and absurd to dark and deceptive have flooded Brazil's internet ahead of Sunday's presidential election. The content has turned Latin America's largest democracy into a real-time laboratory for electoral use of artificial intelligence.

Brazil, a chronically connected nation of 214 million people, has seen online disinformation define its elections for nearly a decade. The current wave of AI material, created by candidates, strategists, and ordinary citizens, now tests the country's ability to monitor online content.

“We made the rules, now we are watching what happens.”

Authorities have issued strict rules governing AI use during elections, but these safeguards had barely been tested until now. Electoral justice decisions on removing publications and suspending channels have sparked legal clashes and accusations of censorship.

In the final stretch, concerns about false information reached the debates. After the confrontation between Lula and Flávio Bolsonaro on Band, social media mobilization urged Globo to adopt a real-time fact-checking system for its debate, a position also publicly defended by the Workers' Party campaign.

In July, Flávio Bolsonaro's campaign displayed an AI-produced video at the PL national convention showing Jair Bolsonaro, barred from campaigning, declaring support for his son. The recreation was challenged in electoral justice and exposed the limits of technologies capable of simulating the presence and statements of real people.

The election has been marked not only by the circulation of disinformation but also by its influence on campaign strategies and conflicts over the limits of electoral justice and digital platforms.