DistantNews
Support us

Brazil’s electoral court majority shields Flávio Bolsonaro from fine over AI-generated Bolsonaro video

From Folha de S.Paulo · () Portuguese

Translated from Portuguese and summarized by DistantNews. Read the original for the full story.

At a glance

News Named sources In the courts
  • Brazil’s Superior Electoral Court formed a majority to reject a fine against Flávio Bolsonaro’s campaign over an AI-generated avatar of Jair Bolsonaro used at a party convention.
  • The court indicated that restrictions on deepfakes would apply when the technology is used in early campaigning, while continuing to debate penalties during the campaign period.
  • The ruling considered whether the avatar’s message requested votes, with the court’s majority finding that it expressed support without directly urging voters to back Flávio Bolsonaro.

Brazil’s Superior Electoral Court has formed a majority to spare presidential candidate Flávio Bolsonaro’s campaign from punishment over an AI-generated avatar of his father, Jair Bolsonaro, shown at a Liberal Party convention in July.

The court also set out a narrower interpretation of when this technology should be barred. It said the prohibition would apply when deepfakes are used in advance campaigning. The full court is considering the issue and discussing a legal position on penalties for deepfakes during the campaign period.

Four justices supported the position, including court president Kassio Nunes Marques, André Mendonça, Dias Toffoli and Antonio Carlos. Under the majority view, the content does not need to be fined or removed after the event because the avatar did not directly ask voters to support Flávio.

Nunes Marques, the rapporteur, said: “The expression does not contain a request for votes. The conclusion is unchanged by the closing phrase, although it refers to the office being sought, because it does not call on citizens.” He said the legal debate was relevant because election law allows expressions of support.

A TSE advertising resolution issued this year prohibits deepfakes that harm or favor a candidate, even when the person reproduced in the video has authorized it. Violations can constitute abuse of political power or misuse of social communications media and may lead to cancellation of a candidate’s registration. Nunes Marques had earlier suggested that AI campaign use could be lawful if it did not harm someone, but later questioned whether allowing supposedly benign deepfakes could trigger excessive litigation.

The expression does not contain a request for votes. The conclusion is unchanged by the closing phrase, although it refers to the office being sought, because it does not call on citizens.

— Kassio Nunes MarquesThe TSE rapporteur explained why he believed the AI-generated avatar did not amount to an explicit request for votes.
About this summary

Originally published by Folha de S.Paulo in Portuguese. Translated, summarized, and contextualized automatically by DistantNews, with a note on how the source frames the story. Not individually reviewed before publishing. How this works.