How AI-Generated ‘Bolsonaro’ Videos Are Shaking Up the 2026 Presidential Race
Weeks before Brazil’s official 2026 presidential campaign period begins, a wave of deepfake videos featuring former president Jair Bolsonaro is sweeping social media. In some clips, a tearful “Bolsonaro” begs for support for his son Flávio, the Liberal Party’s (PL) presidential candidate; in others, the AI-generated figure denounces Flávio as a “puppet” and urges votes for incumbent Luiz Inácio Lula da Silva. The fabricated videos, including one shown at Flávio’s own nomination convention, have forced the country’s electoral authorities and the Supreme Court to confront a technology that blurs the line between political strategy and illegal propaganda.
The controversy escalated on 28 July, when Supreme Court Justice Alexandre de Moraes gave Bolsonaro’s defense 48 hours to explain whether the ex-president authorized the use of his image and voice in the convention footage. Bolsonaro is currently serving a criminal sentence under house arrest for a coup attempt, is barred from social media, and has his visitation rights partially suspended. His lawyers stated he did not consent to the video, noting that Flávio has been prohibited from visiting him for 90 days. Moraes warned that any proven cooperation could reverse Bolsonaro’s humanitarian house arrest to closed prison.
The Brazilian electoral court (TSE) already prohibits deepfakes in electoral propaganda, regardless of whether they harm or benefit a candidacy. However, lawyers are divided on whether the ban applies during the pre-campaign phase that runs until 16 August. While one expert argues the current rules are ambiguous before official campaigning begins, another contends that settled jurisprudence extends all formal propaganda rules to the pre-election period. The PT-led federation has already filed a complaint against Flávio and the PL for early propaganda and irregular use of AI, with TSE President Kassio Nunes Marques assigned as rapporteur.
Where Brazil’s Electoral Justice Draws the Line on Deepfakes
The Legal Gray Zone: Pre‑Campaign vs. Official Propaganda
The TSE’s 2019 resolution (No. 23,610) unequivocally bans deepfakes in electoral advertising, but its wording leaves room for debate before the formal campaign period. Alberto Rollo, a São Paulo lawyer, argues that the prohibition is “clear and objective” only for propaganda once the window opens, leaving the pre‑campaign phase in doubt. Fernando Neisser of FGV/SP counters that precedent firmly applies all formal advertising rules year‑round — including technology bans — making the convention video already illegal. This disagreement matters because much of the AI content circulating on TikTok and other platforms was posted weeks before 16 August, potentially escaping clear sanctions if the stricter view prevails.
Enforcement Realities: Case‑by‑Case Battles and Platform Cooperation
Even if courts deem deepfakes illegal, the sheer volume of videos creates an enforcement challenge. Rollo predicts a flood of cases that will force the TSE to analyze each one individually, warning that “something will slip through.” The court relies on public reporting through a disinformation alert system and has signed cooperation agreements with major social media platforms to detect and remove deceptive AI content. Yet as Neisser notes, campaigns themselves will likely pick only the most visible cases to litigate, leaving lower‑visibility deepfakes to circulate. The TSE declined to comment on pending cases, and its capacity to police every AI‑manipulated video remains untested.
Broader Consequences for the Bolsonaro Family
The personal stakes for Jair and Flávio Bolsonaro are high. If a court finds that Flávio’s campaign knowingly used the deepfake, it could face fines, content takedowns, and — if repeated — eventual charges of abuse of political power, which can nullify a candidacy and mandate. For the elder Bolsonaro, any evidence he collaborated would breach his social‑media ban and house‑arrest terms, potentially sending him back to a high‑security prison. The defense’s denial of authorization is a critical, yet unverified, claim that the Supreme Court is now examining.
Risk & Opportunity Assessment
| Commercial Risk | Low | No directly named commercial entities are involved; the dispute centers on political campaigns and electoral law. |
| Competitive Risk | Low | Competitive dynamics among political parties exist but are not a corporate risk. |
| Regulatory Risk | High | Flávio Bolsonaro’s campaign faces an active TSE complaint for early propaganda and AI misuse, plus Supreme Court scrutiny that could trigger fines, content removal, and — if repeated — charges of abuse of power leading to loss of candidacy or mandate. |
| Reputation Risk | High | The spread of deepfakes — both favorable and hostile — risks associating the Bolsonaro name with electoral manipulation, and a court finding that the ex‑president cooperated would severely damage both his and his son’s public standing. |
| Technology Disruption | High | The election is demonstrating that deepfake technology can now be deployed at scale to mimic a banned politician, posing a direct threat to the integrity of campaign messaging and the enforceability of electoral bans. |
| Commercial Opportunity | Low | This specific situation does not present a clear business opportunity; it is a test of legal and institutional resilience. |
Comments 0