Brazil’s election enters the age of synthetic politics

Two days before Brazilians head to the polls for the first round of the presidential election, the campaign has acquired a distinctly artificial flavour. The problem is not that politics has become less theatrical; it is that the theatre now comes with software. A deepfake avatar of former president Jair Bolsonaro has been used to support his son’s candidacy, while an AI-generated working-class woman appears in another piece of content attacking President Luiz Inacio Lula da Silva.

This is not mere online mischief. According to the Brazilian intelligence agency, disinformation is set to be one of the main tests of this election. That is hardly a comforting diagnosis, but it is at least an honest one. In an environment where voters are expected to sort fact from fabrication at speed, the old campaign tricks look almost quaint.

The timing matters. With the first round only days away, synthetic images and voices can travel faster than corrections, and by the time reality catches up, the damage is already done. The Bolsonaro avatar and the fabricated female voter are not isolated curiosities; they are examples of how political messaging now borrows the tools of deception and dresses them up as persuasion.

Brazil’s race is therefore about more than the usual contest between candidates and camps. It is also about whether the public can trust what it sees, and whether elections can survive being flooded with content that looks human enough to be believed. The country’s intelligence service has put the issue plainly. Campaigns may still be fought by people, but increasingly they are amplified by machines that have no stake in truth and every interest in confusion.

That is the uncomfortable lesson from Brazil’s election week: democracy now has to compete not only with rivals, but with replicas. And replicas, unlike politicians, do not tire, blush or apologise.

Written by Freya Stensrud freya.stensrud@alpineweekly.com