
The information displayed in the AIM should not be reported as representing the official views of the OECD or of its member countries.
Meta's AI, integrated with WhatsApp, refused to generate campaign images for President Lula (PT) with the number 13, while allowing similar images for Flávio Bolsonaro (PL, 22) and others. The system also inserted Bolsonaro's number in unrelated requests, raising concerns about political bias and electoral fairness in Brazil.[AI generated]
Why's our monitor labelling this an incident or hazard?
The AI system is explicitly involved as it generated political campaign images with biased content (number 22 inserted without request and refusal to use number 13). This behavior is a malfunction or misuse of the AI system's outputs. The harm is a violation of legal obligations for equal treatment in elections and potentially a breach of rights to fair political participation, which falls under violations of human rights or legal obligations. The incident is materialized, not just a potential risk, as the biased outputs have already occurred and have been subject to official complaints and legal rulings. Hence, this qualifies as an AI Incident rather than a hazard or complementary information.[AI generated]