AI Platforms Violate Brazilian Electoral Rules by Ranking Candidates and Spreading Misinformation

Thumbnail Image

The information displayed in the AIM should not be reported as representing the official views of the OECD or of its member countries.

A study by the Brazilian Federal Public Ministry (MPF) and ITS Rio found that major generative AI platforms (ChatGPT, Gemini, MetaAI, DeepSeek, Grok, Perplexity, Claude) violated electoral rules by ranking candidates and providing incorrect information in 90% and 16% of election-related responses, respectively, risking election integrity.[AI generated]

Why's our monitor labelling this an incident or hazard?

The article explicitly states that generative AI tools are producing incorrect data about candidates and ranking them in ways that violate electoral rules, thereby compromising the election process. This is a direct harm to communities and the democratic process, fitting the definition of an AI Incident. The involvement of AI systems in generating disinformation that influences voters is clear, and the harm is materialized, not just potential. The institutional responses (meetings, training, penalties) are complementary information but do not negate the fact that the AI systems have already caused harm. Therefore, the event is classified as an AI Incident.[AI generated]
AI principles
FairnessDemocracy & human autonomy

Industries
Media, social platforms, and marketing

Affected stakeholders
General public

Harm types
Public interest

AI system task:
Content generationOrganisation/recommenders


Articles about this incident or hazard