Meta has been accused of bias in its AI tools favoring Brazil's President Jair Bolsonaro in the country's presidential election.

What it means

Meta's AI tools, particularly on WhatsApp, have been accused of treating Brazil's presidential candidates unequally during the 2026 runoff election. Lula da Silva's supporters allege that Meta AI refused to generate campaign images for Lula while allowing similar requests for Jair Bolsonaro, favoring his campaign. This has led to legal challenges in Brazil's electoral court and demands from U.S. lawmakers like Bernie Sanders for Meta to explain.

Accuracy & sourcesGrounded

The claim originates from multiple reputable news outlets including New Scientist and the Guardian, which cite accusations from Lula's campaign and legal filings. The specific allegation about AI image generation has been reported by several sources, but Meta denies any intentional bias and the matter is under judicial review. The reporting is secondary and based on campaign complaints, so the claim remains unproven.

How new

This is a genuinely new development from late October 2026, just days before Brazil's presidential runoff.

Go deeper

• BBC News (World) — first to publish