Brazil's election authority sets AI compliance deadline for big‑tech ahead of 2026 vote
The Superior Electoral Court (TSE) issued Portaria No. 463/2026, giving digital platforms with more than five million monthly active users in Brazil until 16 August 2026 to submit AI‑compliance plans. The plans must detail safeguards to prevent AI‑generated content that recommends candidates or includes non‑consensual sexual material involving candidates, and to enable removal of illicit political content.
Big‑tech firms—including Google, Meta, OpenAI, X and TikTok—are required to present these measures, and the TSE signed an unprecedented agreement with Google, Meta and OpenAI to identify AI‑generated misinformation, with fines of up to R$ 30 000 for violations.
Despite the new rules, a study by the Institute of Technology and Society of Rio (ITS Rio) found that four of the seven leading AI platforms (ChatGPT, Gemini, Meta AI and DeepSeek) continued to rank candidates when asked election‑related questions, contravening the TSE’s Resolution 23.755/2026.
Minister Kassio Nunes Marques, TSE president, was appointed rapporteur of a lawsuit alleging a coordinated network of Instagram and TikTok profiles that use AI to produce large‑scale attacks on President Luiz Inácio Lula da Silva. The court can order technical audits and request additional data from platforms under the new regulations.
Entities: Google · Kassio Nunes Marques · Meta · OpenAI · Tribunal Superior Eleitoral · Tribunal Superior Eleitoral (TSE)
Claims
What the coverage asserts, and how well corroborated each claim is across sources.
- [● 2 SOURCES] President Kassio Nunes Marques will review, approve, or revoke platform accreditation based on compliance plans. (Portaria and TSE statements)
- [● 2 SOURCES] Platforms with more than 5 million monthly active users in Brazil must submit detailed compliance plans; smaller platforms may need simplified plans. (required)
- [● 3 SOURCES] Four of the seven major AI platforms – ChatGPT, Gemini, Meta AI and DeepSeek – continued to rank candidates after the TSE prohibited such recommendations. (ITS Rio research)
- [● 3 SOURCES] TSE issued Portaria No. 463/2026 requiring digital platforms to submit AI‑compliance plans by 16 August 2026. (Portaria No. 463/2026)
- [● 4 SOURCES] The TSE can order technical audits and request additional data from platforms under the new regulations. (Portaria No. 463/2026)
- [○ 1 SOURCE] The Public Security Test (Teste Público de Segurança) has been conducted eight times, allowing citizens to propose attacks on the voting system. (TSE description of TPS)
- [○ 1 SOURCE] TSE signed an agreement with Google, Meta and OpenAI to identify AI‑generated content and impose fines up to R$ 30 000 for violations. (TSE‑Google‑Meta‑OpenAI agreement)
- [○ 1 SOURCE] Brazil’s electronic voting machines undergo about 40 audit steps and have no proven fraud cases since 1996. (TSE spokesperson Júlio Valente)
- [● 2 SOURCES] Minister Kassio Nunes Marques was appointed rapporteur of a lawsuit alleging coordinated Instagram and TikTok profiles using AI to attack President Luiz Inácio Lula da Silva. (court filing)
- [● 2 SOURCES] The lawsuit alleges more than 30 coordinated profiles with about 1.46 million followers and over 23 000 AI‑generated publications targeting Lula and the PT. (required)
- [● 2 SOURCES] Compliance plans must detail safeguards to prevent AI‑generated content that recommends candidates or includes non‑consensual sexual material involving candidates. (Portaria No. 463/2026)