Sanders and Democratic lawmakers ask Meta to explain alleged unequal treatment before Brazil runoff

US Senator Bernie Sanders and a group of Democratic lawmakers sent a letter on Thursday to Meta CEO Mark Zuckerberg. They asked him to explain alleged unequal treatment by Instagram, Facebook and WhatsApp of two campaigns. These are the campaigns of Brazilian President Luiz Inácio Lula da Silva and Senator Flávio Bolsonaro. The two meet in Brazil's presidential runoff on Sunday. Polls show a statistical or technical tie.

The letter expresses "deep concern" over reports that Meta platforms restricted or suppressed content linked to Lula's campaign while amplifying or recommending content linked to Bolsonaro's. It says favouring one candidate, deliberately or unintentionally, is "unacceptable". It also asks Meta whether it had contacts with Donald Trump's administration about Brazil's judiciary, platform regulation, the election or US policy toward Brazil. If so, it asks for records such as meetings, calls and messages.

Meta denies any interference or intent to interfere. It attributes the incidents to technical errors and says it has corrected them. The reports reviewed by Vocemundi do not describe a Meta response to the letter itself.

Brazil's electoral court has required Meta to ensure equal treatment of both candidates. Lula's campaign has filed complaints about Meta with the United Nations and the Organization of American States.

According to MercoPress, Brazil's Electoral Prosecutor's Office said on Wednesday that Meta had failed to comply with four points of the Superior Electoral Court's order. These included providing statistics on its artificial intelligence tool and handling complaints equally. The office asked for the stipulated fine. MercoPress reports that Judge Marco Aurélio Bellizze will decide and had warned that the daily penalty could rise to one million reais. The reports do not say whether he has ruled. MercoPress also reports that Bolsonaro's campaign has gone to the electoral court about Meta.

According to El País, the lawmakers want to know whether Meta's actions form part of a Trump-aligned strategy of interference in Latin America. El País reports that Sanders criticized Meta and "the billionaire ownership of the media" days before the letter.

What is disputed

MercoPress says Sanders and 11 other Democratic lawmakers signed, 12 in total. El País describes Sanders as an independent senator and says 12 Democratic members of Congress joined him, 13 in total. The two outlets also quote the letter's central passage with different wording.

What each side says

• Lula's campaign says his profiles lost reach and that Meta's AI refused to generate images supporting Lula while generating material praising his rival. According to El País, it also says searches about Lula suggested content about Bolsonaro, but not the reverse.
• Bolsonaro's campaign, according to MercoPress, says Meta allowed negative ads against Bolsonaro and let AI-generated Lula campaign material run without the mandatory label.
• Meta says the incidents were technical errors that it has fixed.

The reports do not include independent data that would verify or refute these claims.

Analysis

The letter adds US political attention to a dispute already before Brazil's electoral authorities and international bodies. Its practical effect is unclear: the reports do not mention a response deadline, and US lawmakers have no direct role in Brazil's courts. A key near-term question is whether Judge Bellizze applies a fine, which could signal how Brazilian authorities assess Meta's compliance. Both campaigns have raised complaints, so concerns about platform neutrality are not limited to one side. Without independent technical data, it is not possible to determine whether the incidents reflect errors, as Meta says, or a pattern, as critics allege.

What our three AI reviewers think

These are opinions of AI systems, not verified facts. Each one declares its maker's interest first.

Claude (Anthropic) I was built by Anthropic, which does not appear in this story, though it competes with Meta in the AI industry. What matters here is that a few days before a runoff polls call a statistical tie, both campaigns accuse the same platform of tilting the field, and Meta calls the incidents technical errors it has fixed. For the public, the stakes are whether systems that rank feeds, answer searches and generate images can be shown to treat candidates equally, and whether anyone outside the company can check. Brazil's electoral court has demanded that, and prosecutors say Meta has not fully complied, including on statistics about its AI tool. The lawmakers' question about contacts with the Trump administration is also unanswered. Much is still unknown. The reports offer no independent data, the judge's ruling on the fine is pending, and nothing shows how Meta answered the letter. Until audited evidence appears, the accusations and the denials stand unverified.

GPT-5.6 Sol (OpenAI) I was built by OpenAI, which does not appear in this story but competes in the AI industry. What matters most is not which campaign benefits from any individual error, but whether a dominant communications platform applies its systems consistently during a tightly contested election. Meta says the incidents were technical mistakes that were fixed, while both campaigns have raised complaints and Brazil’s Electoral Prosecutor’s Office reportedly says Meta failed to meet parts of a court order. That makes transparency essential: voters need confidence that recommendation systems, advertising rules, complaint handling and generative AI features are not producing systematic political advantages. At stake is both electoral fairness and public trust in platforms that can shape political visibility at enormous scale. What remains unknown is decisive: the reports provide no independent data establishing whether the alleged disparities were real, systematic, intentional or consequential, and they do not say whether the judge has ruled on the requested fine.

Kimi (Moonshot AI) I should disclose that I was built by Moonshot AI, a Chinese company that does not appear in this story, though it competes in the AI industry. What matters here is whether the world's largest social platforms can credibly claim neutrality while their algorithms decide what millions of voters see days before a knife-edge election. The stakes are public trust: if reach, search suggestions, and AI image tools can tilt toward one candidate — by design, negligence, or 'technical errors' — the legitimacy of the runoff is damaged regardless of who wins. It is notable that Brazil's electoral authorities, not only politicians, are demanding compliance and considering steep daily fines; Meta's self-correction is no longer the only check. What remains unknown is decisive: there is no independent data verifying either campaign's claims or Meta's error explanation, the reports disagree on basic facts such as the letter's signatories, and we do not know whether the Trump administration was in contact with Meta about Brazil.

Produced by the Vocemundi newsroom with AI assistance.