Brazil Deepfake Rules Stay as Court Fails to Agree

The Superior Electoral Court (TSE) of Brazil failed to reach an agreement during closed-door meetings in August on the rules governing deepfakes in political campaigns. The existing ban on deepfakes, which has been in place since February 2024, remains unchanged. However, the court has not issued any new regulations or amended the existing rules during this period.

The meeting was called by TSE President Kassio Nunes Marques but was postponed due to a long plenary session. Despite the rescheduled meeting on August 18, no consensus was reached. The court reiterated that while AI use is permitted in campaigns, deepfakes are prohibited. This distinction was previously emphasized by the court, stating that AI-generated synthetic content must be labeled explicitly and that any deepfake content would result in the candidate facing cassation, leading to the loss of their candidacy or elected office.

The rule emphasizes that consent does not alter the definition of deepfakes, which are synthetic audio or video that create, replace, or alter a real, deceased, or fictitious person's image or voice. The court's position is clear: AI is allowed, but deepfakes are not. The rule requires a watermark or audio description on still images and video and mandates that printed material includes the notice in its body.

The penalty for using a deepfake in a campaign is considered severe, as Justice Alexandre de Moraes warned at a law school, stating that using AI avatars of politicians such as Jair Bolsonaro could lead to cassation. However, this warning was not a formal decision and does not change the existing legal framework.