Submit incident
Documented

Brazil's 2026 Campaigns Flooded Voters With Unlabeled AI Content the Law Was Supposed to Stop

September 15, 2026
Curated by Team Raidu · Reviewed by Shiva Ganesh
oecd:2026-09-15-d7d1View source ↗
LinkedInX

What happened

Brazil's electoral law already required campaigns to label AI-generated content. That requirement did not matter much in practice. A study conducted by Data Privacy Brasil and Aláfia Lab, covering the country's 2026 pre-campaign period, found that nearly two-thirds of AI-generated political content circulating online carried none of the disclosures the law demanded.

The study covered content shared during the period leading up to the formal campaign season, when parties and candidates were actively shaping public opinion ahead of the election. Researchers identified material that included deepfakes and deliberate disinformation alongside conventional campaign posts, all of it generated by AI tools and none of it labeled as such. The legal disclosure requirement exists precisely because voters cannot reliably distinguish synthetic content from real on their own. The study shows that, in the absence of enforcement, the law functioned as an aspirational standard rather than a binding one.

The pattern was not confined to anonymous accounts. Prominent figures, including Flávio Bolsonaro and the Liberal Party (PL), were among those sharing unlabeled AI material. That detail matters because it rules out the explanation that labeling failures were driven mainly by low-sophistication actors who did not know the rules. The campaigns that produced and distributed this content understood what the law required and did it anyway.

What the study documents is a compliance gap that operated largely in the open. Deepfakes and synthetic campaign content do not hide from view the way a backdoor in a software system does. They circulate in public, reach large audiences, and shape voter perception in real time. The failure here is not that the content was hard to find but that no one was positioned to confirm its origin at the moment it was released, before it was shared thousands of times across a national audience.

The deeper problem the incident reveals is the absence of any record that could establish, after the fact, what system produced a given post, who authorized its release, and whether the required disclosure was present at publication. Electoral accountability depends on that kind of provable record of what a system did, who controlled it, and when. Without it, a post that violates labeling law looks identical to one that complies, and the difference only surfaces if a researcher examines a sample months later, well after the audience has moved on.

Reported impact

Affected parties
Not publicly disclosed
Harm type
Not publicly disclosed
Scale
Not publicly disclosed
Financial impact
Not publicly disclosed
Regulatory action
Not publicly disclosed

Classification

Organization
Not publicly disclosed
AI system
Not publicly disclosed
Industry
Not publicly disclosed
Country
Not publicly disclosed
Provider
Not publicly disclosed
Incident type
Not publicly disclosed

Relevant governance controls

Governance control mapping is not available for this record.

  • No controls mappedNot publicly disclosed

Control mapping is analytical. It does not state that any control would have prevented the incident.

Sources and evidence

This record was researched and written by the Index. The event is also catalogued in the following database, which is listed for cross-reference.

OECD AI Incidents Monitor
Also catalogued in
Brazil's 2026 Campaigns Flooded Voters With Unlabeled AI Content the Law Was Supposed to Stop
2026-09-15