Meta AI accused of bias in Brazil election — SkimNews
Get the Geopolitics newsletter
Daily geopolitics — wars, elections, sanctions, the diplomatic moves that move markets. Free.
- Lula's campaign formally accused Meta of favoring Flávio Bolsonaro by suspending its advertising account, suppressing posts, and blocking AI-generated content ahead of the 25 October runoff.
- Meta claimed its systems mistakenly suspended Lula’s ad account for one hour and blocked prompts containing 'Lula' and 'Bolsonaro'—but not the first names Luiz or Flávio—denying any intentional bias.
- Electoral prosecutors in Brazil stated Meta failed to fully comply with court orders requiring equal treatment of both presidential candidates during the election.
- United Nations human rights experts expressed concern over possible unequal treatment of the two campaigns by Meta’s platforms, citing the integrity of democratic processes.
- David Nemer at the University of Virginia noted that repeated AI outcomes favoring one candidate cannot easily be dismissed as technical glitches due to the opacity of Meta’s systems.
- Fact-checking organization Lupa identified 920 pieces of AI-generated or manipulated election content before the first round, 60% of which lacked labels, mostly on Facebook and Instagram.
- Heloísa Massaro at InternetLab found AI chatbots often refused factual queries about candidates but would generate misleading claims about election integrity for campaign use.
Why it matters: With 8 million participants in Lula’s WhatsApp-driven 'Army' and widespread unregulated AI content, Meta’s opaque systems now directly influence voter perception in a tight election—where even brief account suspensions or biased outputs could shift momentum. The failure of Meta’s safeguards raises real-time concerns about private tech control over public democratic processes in Brazil and beyond.
Ask SkimNews
