Submit incident
Documented

Deepfake Videos of Argentina's Former President Tried to Steal an Election in Its Final Hours

January 1, 2025
Curated by Team Raidu · Reviewed by Shiva Ganesh
aiaaic:AIAAIC1991View source ↗
LinkedInX

What happened

In the closing hours of Buenos Aires' May 2025 legislative election, two videos began circulating on X that showed former Argentine President Mauricio Macri appearing to endorse rival candidate Manuel Adorni and, separately, announcing that his own party was withdrawing its candidate from the race. Neither thing had happened. Both videos were fabrications, and by the time they were widely recognized as fakes, they had already moved through thousands of accounts and reached voters still deciding how to cast a ballot.

The videos targeted one of the highest-profile figures in Argentine opposition politics and were designed to produce a specific effect: confusion about who the PRO party was actually backing and whether its candidate, Silvia Lospennato, remained in the race. The PRO party denied both claims and filed legal action, naming Milei's La Libertad Avanza party as a beneficiary of the deception. Pro-government accounts on X amplified the videos, accelerating their reach during the hours when campaigns have the least legal room to respond.

Argentina's electoral court intervened and labeled the incident an attempt at digital fraud. Macri described the videos as "a direct attack on democracy." Experts who reviewed the episode warned that this kind of deepfake, deployed in the final hours before polls close, has a specific and calculated advantage: it is designed to land after the normal response window has closed, so the damage runs through the vote count before corrections can reach the same audience the fake did.

What made the episode possible was less any novel technical sophistication and more the cost of entry. Generative AI tools capable of producing convincing video fabrications of named public figures are cheap and widely available. Argentina's polarized political environment gave those tools a ready motive. The slow response from social media platforms gave them time. None of those conditions required a state actor or a specialized operation. They required a campaign cycle, a target, and an afternoon.

The deepest failure this incident surfaces is the absence of any mechanism to prove what a video is, where it originated, and when. The electoral court could label the episode fraud, and Macri could call it an attack, but neither institution had access to a provable record of what a system produced, who uploaded it, and how the platforms handled the flag. Without that chain of evidence, takedown arrives after the harm is done, and attribution relies on what accounts happened to amplify the content rather than where it was actually made.

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.

AIAAIC Repository
Also catalogued in
Deepfake Videos of Argentina's Former President Tried to Steal an Election in Its Final Hours
2025