< Back to all clusters
[POLITICS] · Brazil · 7 sources

started · updated

Brazil 2026 elections face surge in AI deepfakes

The 2026 Brazilian election campaign is facing a significant surge in AI-generated disinformation. A report by the VigIA project, a collaboration between the Lupa fact-checking agency and Unicamp, found that 554 deepfakes were identified among 920 pieces of content produced or modified with artificial intelligence between August 16 and September 28. These deepfakes represent 60% of the total AI-related content analyzed.

Commonly used tactics include creating videos that simulate journalists delivering false news or celebrities falsely declaring political support. Notable targets of these manipulations include Anitta, Caetano Veloso, Neymar, and Virginia Fonseca. In one instance, Fonseca’s representatives denied a video showing her supporting Flávio Bolsonaro was authentic.

The report highlights that while 29 deepfakes were disseminated by candidates or political parties, the vast majority—525 posts—were created by common or anonymous users, posing a major challenge for electoral oversight. The Superior Electoral Court (TSE) has prohibited deepfakes in propaganda that could deceive voters, noting that violations can lead to the revocation of candidacies or mandates.

In a separate statement, TSE President Kassio Nunes Marques addressed concerns regarding international interference, asserting that democratic nations engage in technical dialogue and knowledge exchange rather than interfering in the sovereign electoral processes of other nations.

Entities

Brazil · Kassio Nunes Marques · Lupa · Tribunal Superior Eleitoral · Unicamp · Universidade Estadual de Campinas · VigIA

Claims

What the coverage asserts, and how many sources carry each claim.