BRASÍLIA — Brazil’s top elections court sat Tuesday to set clearer limits on AI-generated campaign content, weeks before the Oct. 4 first-round presidential vote, as deepfakes and chatbots crowd the race between President Luiz Inácio Lula da Silva and Senator Flávio Bolsonaro.
The case before the TSE involves an AI avatar of former President Jair Bolsonaro, who is under house arrest and barred from addressing the public, shown at the Liberal Party convention that launched his son’s ticket. Lula’s campaign has asked the court to treat the video as a prohibited deepfake. Reuters said Tuesday’s ruling was expected to sharpen the definition of what counts.
TSE rules already require labels on AI-generated material, ban deepfakes used to help or hurt a candidacy, bar generative systems such as ChatGPT from ranking candidates, and black out new AI political content immediately before and after voting. The court logged 19 AI cases in the first 11 days of this campaign. In early August, Justice André Mendonça ordered removal of a video showing the elder Bolsonaro with banker Daniel Vorcaro in scenes that never happened.
A Projeto Brief survey in May found nearly 63 percent of voters willing to consult AI tools about candidates. A study by ITS Rio found seven widely used chatbots still ranked or ordered candidates in 90 percent of answers, despite the ban. João Victor Archegas, legal coordinator at ITS Rio, said the problem is not the tool but “its potential use in certain contexts that may lead to dangerous outcomes for our democracy.”
Sources: Reuters via The Straits Times.
