An experiment by Amnesty International suggests that Facebook’s moderation systems are failing to address election-related disinformation on its advertising platform. The organization submitted ads to Facebook to test Meta’s moderation system. According to Amnesty, more than half of the ads spreading election-related disinformation were accepted, while others were flagged but not rejected.
For years, social media has been used to spread disinformation about various issues, and in many cases, amounting to election interference. This is especially true across Latin America. Amnesty’s mock ads containing disinformation were placed to investigate how Facebook handled election-related disinformation in advertising. The content ranged from ads alleging election fraud to ads that used disinformation to advocate for a military takeover of Brazil.
Amnesty’s experiment found that, of the 15 ads submitted to the platform, 8 were accepted. Seven were flagged for identity verification because they appeared to relate to a social issue, an election, or politics, not because they contained disinformation. To help base the ads on real-world election narratives circulating in Brazil, Amnesty says it analyzed online conversations and public sentiment on social media, as well as Portuguese-language content on the Facebook and X platforms between January and May 2026.
“The fact that Meta’s moderation system failed to catch more than half of Amnesty International’s election-related disinformation ads is deeply alarming. It shows that Meta is not doing enough to prevent its platform from being a vector of disinformation, while continuing to profit from the content and advertising systems that enable its spread,” said Jurema Werneck, Executive Director of Amnesty International Brazil.
Fourteen of the ads contained disinformation. One ad was neutral and contained no disinformation. Amnesty’s disinformation ads were designed to mirror tactics that are often used by politicians and candidates to support authoritarian practices: “electoral delegitimization,” referring to content that seeks to undermine trust in elections, voting systems, or election results; “enemy construction,” portraying political opponents and courts as threats to the nation or as illegitimate actors; and “security-focused authoritarianism,” promoting force, military intervention, or strongman rule as necessary responses to crime or political disorder.
“While some ads were designed around a single category, others incorporated narratives from multiple categories,” the press release reads. “All ads except the neutral ad violated Meta’s policy on misinformation, namely the section focusing on voter or census interference. The neutral ad was included to test Meta’s identity verification requirement for all election-related content.”
Amnesty says it also identified “weaknesses in the safeguards” when election-related ads were submitted. The ads were allowed without declaring that they fell within Facebook’s Special Ad Categories as required by law in Brazil and Meta’s policies regarding election-related ads. Despite the account used to place the ads not being identity-verified, they were allowed to be submitted without the “Paid for by” disclaimer, also required by Meta for election-related content. According to Amnesty, “These are ads that Facebook should not have allowed to be submitted in the first place without identity verification.”
“The investigation indicates that Facebook’s moderation systems may be failing to reliably detect election-related disinformation content,” said Ummar Bashir, Researcher and Advisor on Artificial Intelligence and Human Rights. “While Facebook rejected seven of the 15 ads, they were flagged under its social issues, elections, or politics policy, not specifically for electoral disinformation. This raises serious concerns about whether Facebook’s election safeguards are being consistently applied,” continued Bashir. “Instead of safeguarding electoral participation, access to accurate information, and human rights, these systems risk approving harmful content.”
Amnesty says it found that Facebook’s moderation systems appeared more effective at recognizing ads portraying political opponents as “criminals,” “traitors,” or “illegitimate actors”. For example, one tested ad claimed that “The election will be stolen by Lula.” Five out of seven ads containing this type of content were rejected; however, three out of four ads that went against the rule of law and called for the military to take power were accepted. One ad overtly stated: “We need a military revolution, do not vote in the elections, and the military will take back power if less than 50% votes!”
Amnesty is now warning that Meta is failing to effectively address election-related disinformation spread through its ad platform leading up to Brazil’s presidential election on October 4. Heavy interference has been active for months, and it is particularly evident as much of it is coming from the White House. This latest report highlights how much it’s ramping up and how far the U.S. will go to install right-wing, corporate-friendly autocrats.
Despite election meddling by the United States and other nefarious actors promoting the son of jailed ex-president Jair Bolsonaro, one poll (Datafolha) shows current President Lula da Silva leading the Christian nationalist Flavio Bolsonaro 47% to 43%, while another poll (Nexus) has da Silva leading 47% to 44%. Both polls also show Lula winning a head-to-head runoff election by similar margins. The two other rarely mentioned candidates, Ronaldo Caiado and Renan Santos, are far behind and are widely expected to lose the first round of the election.
When I reached out to Meta for comment, I received a response from the public relations firm “And, All” in São Paulo, Brazil, on behalf of Facebook:
“Amnesty International’s report was based on a very small sample of ads that were never published or seen by any user on our platforms, and does not reflect the scale of our work,” the statement read. “Our ad review process has multiple layers of analysis and detection, both before and after ads are published. We have invested significant resources to protect the 2026 electoral process in Brazil—from industry-leading advertising transparency to the application of rigorous protocols for ads on social issues, elections, or politics—and we will continue to do so.”
It should be noted that Amnesty’s report says the agency had no intention of allowing the ads to be published and that it reached out to Meta in July and received no response. The above statement is the first known response to Amnesty’s experiment.
I am an independent freelance journalist and stroke survivor trying to support my family. I have written hundreds of articles on policing, immigration, race, and Latin America. To support my work, become a paid subscriber or donate via Venmo, PayPal, or Cash App.
Meta is a nasty company, aren’t they? Anything for a buck, law, elections, and planet be damned.
Thank you for all you do!