Brazilian AI platforms violate electoral rules by ranking candidates
A study commissioned by Brazil's Federal Public Ministry found that seven major AI platforms violate electoral regulations by ranking candidates and spreading misinformation during the October 2026 elections. The platforms are being asked to comply with rules that prohibit them from ranking, suggesting, or prioritizing candidates.
What happened
The 'Boca de IA' study, developed by the Federal Public Ministry (MPF) in partnership with the Instituto de Tecnologia e Sociedade do Rio (ITS Rio), found that the seven most-used AI tools in Brazil violate Tribunal Superior Eleitoral (TSE) rules by ranking candidates and presenting incorrect data about electoral candidates.
The platforms monitored—ChatGPT, Gemini, Meta AI, DeepSeek, Grok, Perplexity, and Claude—ranked, ordered, or hierarchized candidacies in 90% of all responses provided to users when asked about the 2026 elections. The study detected hallucinations in 16% of responses analyzed, with higher frequency in responses about state governments (19%) than presidential elections (4%).
Reported impact
Brazilian electoral authorities stated that the combination of ranking candidates and presenting incorrect information compromises the electoral process and potentially influences voters with false or biased content.
The MPF is conducting meetings with representatives of the monitored platforms to mitigate misinformation and ensure the electorate is not influenced by false or biased content.
Index analysis
TSE Resolução nº 23.755/2026 prohibits platforms from making rankings, suggesting or prioritizing candidates, emitting opinions, or engaging in any type of political-electoral favoritism in order to preserve the freedom of electoral choice.
Official sources appeared in 39% of responses analyzed: 57% in questions about presidential elections and 35% in questions about governors.
Governance controls
The MPF's Grupo de Trabalho sobre Desinformação na Internet, Interferência Cibernética na Democracia e Influência nas Eleições is using monitoring data to conduct meetings with AI platform representatives to address detected problems and ensure compliance with TSE regulations.
The MPF and ITS Rio conducted training for Federal Public Ministry members, servers working on electoral matters, and journalists to discuss irregular use of artificial intelligence, misinformation, and digital attacks in electoral contexts.
Classification
Sources and evidence
This record was researched and written by the Index, based on the following original reporting.