AI campaigns in Argentina’s elections

The posters that splash the streets of Buenos Aires have a certain Soviet touch. There was one of Sergio Massa, one of the presidential candidates in Argentina, dressed in a shirt with what appeared to be military medals, pointing to a blue sky. He was surrounded by hundreds of older people – with monotonous outfits, serious and often disfigured faces – who looked at him with hope.
The style was not a mistake. The illustrator had received clear instructions.
“Illustration of a political propaganda poster from Gustav Klutsis with a leader, Massa, standing tall and firm,” said a message that Massa’s campaign introduced into an artificial intelligence program to produce the image. “Symbols of unity and power fill the environment,” the prompt continued. “The image radiates authority and determination.”
Javier Milei, the other candidate in Sunday’s runoff election, has countered by sharing what appear to be artificially created images depicting Massa as a Chinese communist leader and himself as an adorable cartoon lion. They have been viewed more than 30 million times.
Argentine elections have quickly become a testing ground for artificial intelligence in political campaigns, with both candidates and their supporters using technology to alter existing images and videos and create new ones from scratch.
Artificial intelligence has made candidates say things they didn’t say and placed them in famous movies and memes. It has generated campaign posters and sparked debates about whether real videos are effectively real.
The prominent role of artificial intelligence in Argentina’s campaign and the political debate it has sparked underscore the growing prevalence of the technology and demonstrate that, with its increasing power and decreasing cost, it is likely to be a factor in many democratic elections around the world.
Experts compare this moment to the early days of social media, a technology that offers new and tempting tools for politics, as well as unforeseen threats.
Massa’s campaign has created an artificial intelligence system that can create images and videos of many of the key players in the elections – the candidates, the running mates, the political allies – doing a variety of things.
The campaign has used artificial intelligence to portray Massa, the serious center-left Minister of Economy, as strong, fearless, and charismatic, including videos showing him as a soldier in a war, a Ghostbuster, and Indiana Jones, as well as posters evoking Barack Obama’s 2008 “Hope” campaign poster and a cover of The New Yorker.
The campaign has also used the system to portray opposing candidate Milei – an extreme right-wing libertarian economist and television personality known for his outbursts – as unstable, placing him in movies like A Clockwork Orange and Fear and Loathing in Las Vegas.
Much of the content has been clearly fake. But a handful of creations have crossed the line into misinformation. Massa’s campaign produced an ultra-fake video, known as a deepfake, in which Milei explains how a market for human organs would work, something he has said philosophically aligns with his libertarian views.
“Imagine having children and thinking of each of them as a long-term investment. Not in the traditional sense, but thinking about the economic potential of their organs in the future,” says the manipulated image of Milei in the fabricated video, posted by Massa’s campaign on their artificial intelligence Instagram account called IAxlaPatria.
The caption of the post says, “We asked an Artificial Intelligence to help Javier explain the organ trade business, and this happened.”
In an interview, Massa said that the first time he saw what artificial intelligence could do, he was shocked. “I wasn’t mentally prepared for the world I was going to live in,” he said. “It’s a huge challenge, we’re riding on a horse whose tricks we don’t know.”
The New York Times then showed him the deepfake video his campaign had created of Milei talking about human organs. He seemed disturbed. “I don’t agree with that use,” he said.
His spokesperson later emphasized that the post was a joke and was clearly labeled as generated by artificial intelligence. His campaign stated in a statement that their use of the technology is for entertainment and political commentary, not to deceive.
Researchers have long expressed concerns about the effects of AI on elections. The technology has the ability to confuse and deceive voters, create doubts about what is real, and spread misinformation through social media.
For years, these fears were speculative as the technology to produce such fake content was too complicated, expensive, and crude.
“Now we have seen this total explosion of incredibly accessible and increasingly powerful toolkits that have been democratized, and that appreciation has changed radically,” said Henry Ajder, an expert based in England who has advised governments on AI-generated content.
This year, a candidate for mayor of Toronto used dark-toned AI-generated images of homeless people to suggest what Toronto would be like if he wasn’t elected. In the United States, the Republican Party released an AI-created video showing China invading Taiwan and other dystopian scenes to illustrate what would supposedly happen if President Biden won re-election.
And Florida Governor Ron DeSantis’ campaign shared a video showing AI-generated images of Donald Trump embracing Anthony Fauci, the doctor who has become an enemy of the American right for his role as the leader of the national pandemic response.
So far, the AI-generated content shared by the campaigns in Argentina has been labeled to identify its origin or is such an obvious forgery that it is unlikely to deceive even the most gullible voters. Instead, the technology has empowered the ability to create viral content that would have previously required the work of entire teams of graphic designers for days or weeks.
Meta, the company that owns Facebook and Instagram, said this week that it would require political ads to indicate if AI was used. Other unpaid posts on sites that employ such technology, including political ones, would not require such information. The Federal Election Commission in the United States is also considering whether to regulate the use of AI in political propaganda.
The Institute for Strategic Dialogue, a London-based research group that studies internet platforms, signed a letter calling for such regulations. Isabelle Frances-Wright, the group’s Director of Technology and Society, commented that the extensive use of AI in Argentine elections was concerning.
“I definitely think it’s a slippery slope,” she said. “What already looks very real now will only look more real in a year.”
Massa’s campaign said they decided to use AI in an effort to show that Peronism, the 78-year-old political movement that backs Massa, is capable of attracting young voters by surrounding Massa’s image with pop culture and memes.
To achieve this, the campaign’s engineers and artists uploaded photographs of various Argentine political figures to an open-source program called Stable Diffusion to train their AI system to create fake images of those real people. They can now quickly produce an image or video featuring more than a dozen notable Argentine political figures doing practically anything.
During the campaign, Massa’s communication team instructed the artists working with the campaign’s AI about the messages or emotions they wanted to evoke with the images, such as national unity, family values, or fear. The artists then brainstormed ways to insert Massa or Milei, as well as other politicians, into content that evokes movies, memes, and…