Publications /
Opinion
Brazil’s presidential election, which concluded on October 25, 2026, was above all a digital battle. This is nothing new; the impact of social media on elections has been widely analyzed since Barack Obama's 2008 win marked a turning point for digital campaigning worldwide. Almost two decades later, social media is an entirely different breed, with a life of its own – and still very little regulation within electoral politics. The danger, however, is that digital players now also include technology platforms themselves. Platforms do far more than host campaign messages, and their systems help determine which political voices reach voters.
Meta, the parent company of WhatsApp, Instagram and Facebook, has been at the center of this debate in Brazil. Allegations of unequal content moderation and distribution prompted complaints and judicial scrutiny during the campaign. On October 21, independent UN human rights experts raised concerns about reported restrictions on political expression across Meta’s platforms and called for transparent explanations and effective remedies. Their intervention did not establish that Meta had deliberately interfered in the election, but it highlighted new potential risks to political participation.
These concerns also arise in a wider debate about the relationship between technology companies and political power, in parallel to the United States’ policy influence in Latin America. Meta’s US$1 million donation to Donald Trump’s inaugural fund and its January 2025 announcement that it would end third-party fact-checking in the United States warrant scrutiny. Neither development proves coordination with the US administration to influence Brazil’s election, but it raises questions about how countries can protect democratic choice when platforms headquartered abroad exercise substantial influence over their information environments.
Protecting votes and informing voters
Brazil’s electronic voting system has faced repeated attacks since the rise of the far-right in the country, almost a decade ago. Despite that, it combines rapid vote counting with extensive security safeguards. Its security and accessibility measures are designed to protect the vote and facilitate participation, including in the most remote communities. Voting takes place on Sundays, public transportation is free, and biometric identification and numerical and Braille codes for candidates further extend accessibility.
Electronic voting enables rapid counting, with results generally available within hours, and provides opportunities for inspection and auditing at multiple stages. Voting machines operate offline, and safeguards include cryptographic protections, source-code inspection and public security tests. Such a level of transparency allows for increased public trust and engagement in the voting process – or so it should.
Despite this holistic ecosystem of technical democratic safeguards, the Brazilian far-right has frequently contested unfavorable results claiming electoral fraud. Repeated allegations of fraud can undermine public confidence even when they remain unsubstantiated. Brazil’s technical safeguards, however, do not shield voters from attempts to influence public opinion.
Here lies the paradox. Considerable effort goes into securing how votes are cast, processed and counted, while voters remain cyber-vulnerable before that point, exposed to opaque systems that shape the information available to them. Private platforms’ algorithms influence what users see and, just as importantly, what they do not see, leaving narrative control in the hands of a highly unregulated digital environment. Where recommendation systems are difficult to scrutinize and misleading content circulates widely, securing the ballot alone cannot safeguard informed political choice. Beyond the voting system, another vulnerability remains: voters themselves.
Algorithmic influence and electoral oversight
Brazil has introduced rules governing digital campaigning, but regulating political advertising does not fully address the influence of content ranking, reach and recommendation. Paid advertisements can be subject to disclosure and spending rules; the distribution of unpaid content raises different oversight challenges. Platforms retain discretion over reach and visibility, dictated by the algorithm and protected as intellectual property, while outsiders often lack the data needed to assess whether their systems treat political actors consistently. In a close election, the capacity to reach audiences beyond a candidate’s existing supporters can be politically significant.
The online reach of federal deputy Nikolas Ferreira, a close ally of Flávio Bolsonaro, drew particular scrutiny. A Sleeping Giants Brasil study examined the ten most recent videos from each of 30 prominent Instagram accounts. Ferreira’s videos averaged 41 million views, with a median of 38 million, despite an account following of approximately 23 million. Unusually high reach does not by itself demonstrate preferential treatment, but these findings prompted calls for explanations.
Brazil’s Superior Electoral Court (TSE) subsequently required Meta and X to report every six hours on the reach of posts by Ferreira and influencer Pablo Marçal during the final stage of the campaign. It also restricted artificial amplification and recommendations to non-followers. The court stated that the available data alone did not establish improper algorithmic favoritism. Separately, allegations involving campaign advertising, search results and AI-generated images prompted requests for additional technical explanations. Meta denied political favoritism and attributed reported incidents to a series of technical problems. These disputes do not prove deliberate interference, but the broader pattern justifies investigation.
The independent UN experts also referred to information, including from Meta employees, suggesting that accounts associated with Lula's side of the contest had been disproportionately restricted. Meta’s denials and the allegations against it underscore the need for independent assessment, highlighting the need for improved oversight in how tech platforms manage electoral content. Commercial confidentiality can make scrutiny difficult, but should not prevent proportionate review of systems with significant electoral effects. Who decides which political content becomes visible, recommended, restricted or removed? Understanding why one post reaches millions of people while another reaches thousands requires access to evidence about how those decisions are internally made.
Platform accountability
There are serious risks when technology companies become central to democratic debate without corresponding mechanisms of accountability. Platforms have moved from facilitating communication to helping shape the information through which people understand themselves, the world around them, and their own political choices. Equal treatment does not require identical reach for every candidate, but it does require clear rules, consistent enforcement, enforceable transparency and effective avenues of appeal. A public democratic process should not depend on opaque private decisions that affected users cannot meaningfully challenge.
The information environment before polling day therefore deserves attention alongside the security of voting itself. Protecting the final act of casting a ballot cannot compensate for persistent exposure to misleading information or unjustified restrictions on legitimate political expression. Electoral integrity must encompass the conditions under which voters form their judgments, while protecting their freedom to disagree and choose. Today, the digital information structure that shapes voter political preferences and decisions remains mostly privately governed, as if it were detached from the democratic process and not an integral part of it.
WhatsApp illustrates a further dimension of this challenge. Its role in everyday communication in Brazil makes it an important channel for political messages – Statista projected 147.68 million users in Brazil for 2024, placing the country among the app’s largest markets globally. Private messaging and public broadcast channels present different oversight problems and should not be treated as interchangeable. In an October 14 investigation, Aos Fatos identified 17 false or misleading posts published between October 7 and 13 on Tropa do Flávio, a WhatsApp channel linked to Flávio Bolsonaro’s campaign. The channel had more than 1.7 million subscribers at the time.
The implications extend beyond Brazil. In countries where a small number of foreign-based platforms mediate much of public debate, dependence on their systems can severely constrain domestic oversight. Similar vulnerabilities may arise elsewhere in the Global South, even as usage patterns, institutions and regulatory capacities differ substantially. When billions of voters are subject to the sways of platforms designed and governed outside their own countries, whose design and governance may not adequately reflect local contexts, democracy is at risk. Information filtering through boosting or restricting content can be the difference between winning and losing – not just an election, but rights, representation, and a political voice. Digital sovereignty should mean the capacity to protect citizens’ rights and enforce transparent rules.
Ways forward
The Brazilian experience strengthens the case for closer scrutiny of how digital platforms distribute political content. Looking ahead, platforms should provide timely explanations of consequential moderation and recommendation decisions, with sufficient information for competent oversight bodies to assess unequal treatment and order remedies where warranted. Stronger governance is needed during elections and also in everyday public debate, as engagement incentives can reward polarizing content. Without this, we risk becoming a society dominated by algorithms, with voting, our highest expression of citizenship, also conquered by their logic.
Practical responses include access to relevant data for qualified independent researchers, proportionate audits, transparent political advertising and rapid appeals during campaigns. Such measures need safeguards for personal data, freedom of expression and due process. Media literacy can also help voters evaluate political claims without transferring responsibility for platform conduct entirely to individual users.
Those who govern digital platforms exercise substantial influence over political debate, even if they do not determine electoral outcomes directly. Electoral integrity should encompass the wider information environment in which citizens form their preferences, and the act of voting is only the final stage of that process. Protecting it requires institutions that enable people to make informed, independent political decisions rather than leaving those conditions to platforms’ commercial incentives alone. Today, those who control social media and digital presence also exercise substantial influence over political life.
Lula’s reelection does not resolve the questions raised during the campaign. The election instead provides an opportunity to examine the relationship between platform power, domestic institutions and democratic choice. The relevant standard must apply regardless of which party wins: political actors should compete under transparent rules, and allegations of unequal treatment should receive timely, evidence-based scrutiny. This is the very basis of democracy.
Brazil’s experience points to a broader agenda for digital sovereignty: protecting the voter’s journey to the polls, improving oversight of the information environment and holding platforms accountable for demonstrable failures. It also showed that the digital space is democracy's new battleground. The task is to strengthen democratic choice while preserving the openness, privacy and pluralism on which that choice depends. And to move towards a future of digital sovereignty.
