Brazil’s Superior Electoral Court (TSE) unanimously fined Meta R$7.5 million over its handling of court orders during the 2026 presidential runoff involving Lula and Flávio Bolsonaro. The dispute turns on two distinct questions: whether Meta complied adequately with the court, and whether its platforms intentionally distorted the contest.

The court’s answer was firm on the first point. Rapporteur Marco Aurélio Bellizze said Meta’s responses were late and incomplete, with company-set deadlines replacing judicial ones, English-language reports and generic explanations that made oversight difficult. He argued that voters need not only the right to choose, but conditions in which that choice can be formed freely in a digital environment whose distribution systems are difficult to scrutinize.¹

But Bellizze did not find sufficient evidence that Meta deliberately sought to benefit either candidate. The ruling rejected a request to treat the conduct as an affront to the justice system and declined to raise the fine to R$10 million. Meta said it applies the same electoral rules to every candidate and plans to contest what it calls false allegations.¹

Sources aligned with Lula’s side emphasize the underlying disparity allegations. His coalition pointed to reduced reach, problems boosting advertisements and differing engagement, while tests cited in the case found Meta AI refused 98.1% of image requests supporting Lula, compared with 24.7% of requests supporting Flávio Bolsonaro. Meta acknowledged technical failures affecting Lula campaign ads and video-view counts, but said they were not directed at any campaign and that its algorithms do not use ideological criteria.²

Lula’s lawyer, Angelo Ferraro, cast the dispute in broader terms, saying, “Brazil is not anyone’s colony and its president will not be chosen by Silicon Valley or Washington.”³ The court, however, confined its sanction to Meta’s compliance failures—not a finding that it had intentionally interfered in the election.