Brazil’s Superior Electoral Court (TSE) has formed a majority to fine Meta R$7.5 million in a dispute over the handling of election-related content during the presidential race. The decision gives Lula’s coalition a financial ruling against the company, but does not conclude that Meta deliberately favored Flávio Bolsonaro.

Lula’s campaign had argued that Meta applied “algorithmic selectivity,” citing recurring problems boosting its advertisements, reduced reach for posts and different engagement levels between the candidates.¹ It also challenged the conduct of Meta AI: tests submitted to the court allegedly showed the tool refusing requests for favorable images of Lula while responding to comparable prompts about Bolsonaro.²

Meta rejected any attempt to influence the election. It acknowledged technical failures affecting advertisements bought by Lula’s campaign and distortions in video-view counts, but said the incidents were not directed at any candidate and that its algorithms do not use ideological or party-political criteria.¹

The court’s majority adopted the reasoning of Justice Marco Aurélio Bellizze, who had already imposed restrictions on the company during the second round. Before that vote, Meta had been barred from reducing reach, limiting recommendations or changing metrics for either candidate’s political content; violations carried a daily R$500,000 penalty.²

Bellizze said the ruling concerned noncompliance with judicial orders rather than proof of intentional favoritism toward Bolsonaro. In his vote, he warned that global actors such as technology companies can wield forces “the State cannot control and before which the Constitution loses effectiveness.”³ The distinction leaves the underlying allegations contested: the campaign points to a pattern of unequal outcomes, while Meta maintains that operational faults—not political preference—explain the disparities.