Free campaign airtime on radio and TV for the first round of Brazil's 2026 elections begins this Friday (28) without a final word from the Superior Electoral Court (TSE), the body that oversees Brazilian elections, on the limits for using artificial intelligence and deepfakes in the race. According to G1, the court's full bench is expected to rule on the issue next week.
The broadcast window runs Monday through Saturday until October 1, three days before the first round. This Friday's programs feature candidates for the Senate, state deputy and governor. Presidential ads start on Saturday (29). Under the time split calculated by the TSE, the coalition backing President Luiz Inácio Lula da Silva (PT) gets 5 minutes and 31 seconds per block, Senator Flávio Bolsonaro (PL) gets 4 minutes and 20 seconds, Ronaldo Caiado (PSD) gets 2 minutes and 2 seconds, and Augusto Cury (Avante) gets 35 seconds. Romeu Zema (Novo) and Renan Santos (Missão) received no network time because their parties fell short of the performance threshold set by a 2017 constitutional amendment.
Current TSE rules ban deepfakes, meaning videos, images or audio manipulated by AI to simulate the speech or actions of real people, when the tool is used to benefit or harm candidates. On Tuesday (25), TSE vice-president André Mendonça proposed that the court debate a thesis on the reach of that ban for the 2026 elections. The dispute gained traction after the PT filed a case against the PL and Flávio Bolsonaro over an AI-made video shown at the party's national convention in July.
Where the parties stand
G1 asked the five main parties fielding presidential candidates about the issue. The PT said its leaders treat the use of AI in the campaign with caution and will wait for the TSE's position, expected next week, before setting its strategy. The PL did not respond, but Flávio Bolsonaro's campaign did: in its view, electoral law cannot aim to ban technologies, and the TSE resolution, as the campaign reads it, "does not prohibit the use of artificial intelligence in electoral advertising, but disciplines its use and imposes a transparency criterion". The ban would apply, the team said, only when the tool is used to create and spread what it called qualified falsehood.
The Novo party defended the free use of generative AI in campaigns and drew a line between the technology in general and deepfakes. In a statement to G1, the party said: "There is no problem at all in using artificial intelligence in the campaign for the most diverse purposes. Zema himself uses it frequently, including in the satirical series 'Os Intocáveis'. But it is clear that, in the case of deepfakes, without the consent of the person portrayed and with the goal of deceiving the voter, it should not be allowed".
The Missão party said it favors AI use, arguing the technology cuts costs for parties without large structures, and called it an "excess" the rule that bars synthetic content reproducing a person's image or voice even with authorization and labeling. The PSD, which launched Ronaldo Caiado's presidential bid, did not reply to G1.
The next move belongs to the TSE. The full bench is expected to rule next week on the limits for AI use in the campaign, with radio and TV advertising already on the air. Broadcasts run until October 1, and the first round is set for October 4.