What is slopaganda, the political propaganda so dear to Vannacci created with generative AI
No one really thinks that Donald Trump pilots an F-16 or that Roberto Vannacci has ever set foot in the Colosseum dressed as Caesar. The point, however, is not this: slopaganda does not aim to deceive about the facts; it aims to glue two images in the mind of the viewer until they become a single sensation.
The philosopher Michał Klincewicz, who studied the phenomenon alongside his colleague from Macquarie University, Mark Alfano, and independent researcher Amir Ebrahimi Fard, describes it this way: placing Satan next to Trump, America next to evil, and repeating the pairing until it becomes fixed in the mind, even when the viewer knows perfectly well that it is a photomontage. It is a form of persuasion that bypasses reasoning and works on instinct.
The term, already registered by Treccani, is born from the fusion of "slop" (the term in English that refers to the substandard and repetitive material churned out in bulk by image and text generators) and "propaganda". It describes content with a hyper-realistic aesthetic, exaggerated tones, and almost zero production costs, designed to disseminate quickly on social media. The paper signed by Alfano, Klincewicz, and Ebrahimi Fard cites the case of News Corp Australia, which produces about 3,000 "local" articles per week generated with artificial intelligence: a volume unimaginable with human labor alone.
The case of Vannacci in Italy has now gone viral, between the Colosseum and lions. What makes slopaganda different from any other form of political communication in the past are three characteristics that reinforce each other: the scale of production, the speed with which content is generated and disseminated, and the ability to customize it for different audiences. Those who use it no longer need to rely on a graphic studio or a director: just a prompt. The result is a constant flow of images and videos that are hard to verify in their origin, ultimately wearing down citizens' ability to distinguish what actually happened from what was simply generated to impress.
View this post on Instagram A post shared by ROBERTVS VANNACCIVS (@robertus_vannaccius)
In Italy, the most recent example bears the signature of the leader of Futuro Nazionale. The video titled "Il Generale Robertvs Vannaccivs scena colosseo Giuseppus Barbonis" shows the former general in the guise of a Roman emperor, sitting in the Colosseum next to Prime Minister Giorgia Meloni, Senate President Ignazio La Russa, and his partner Camelia Mihailescu. Political opponents, from Elly Schlein to Giuseppe Conte, including Romano Prodi and Laura Boldrini, appear instead tied up, destined to be devoured by lions.
No one passes off those images as real: the artificial rendering is evident from the very first frame. But this is precisely the mechanism, visually associating opponents with defeat and loss of power, without needing to prove anything.
It is not an isolated incident. In December 2025, from circles close to the Vannacci team, the piece "Remigration: the only solution" was already circulating, likely created with the help of generative artificial intelligence. A few months earlier, in March, another video showed the former general riding giant waves on a surfboard with heroic features, while European Commission President Ursula von der Leyen was depicted as a defeated boxer in the ring. The same pattern: strength on one side, weakness on the other.
In the United States, experimentation started earlier and has touched even more explicit tones. President Donald Trump has already shared images of himself in papal robes, in a Superman suit, and in a doctor’s coat. On October 1, 2025, amid negotiations over the federal budget, he published on Truth Social a 35-second deepfake video in which Senator Chuck Schumer, complete with fake mustaches and a computer-generated sombrero next to colleague Hakeem Jeffries, delivers a vulgar monologue full of racial references, with mariachi music playing in the background.
Again, the goal was not to make people believe that Schumer actually said those words, but to associate the political opponent with ridicule, in front of millions of followers.
The research by Alfano, Klincewicz, and Ebrahimi Fard recalls an old idea, that of Francis Bacon, according to which "knowledge is power": those who control the informational environment of a group, in a certain sense, decide in its place. With slopaganda, that control becomes economic and accessible to anyone with access to an image generator. The risk for democratic societies is constant exposure to content whose origin can no longer be verified with certainty. And here the challenge changes nature: the challenge will no longer be just to recognize the false but to resist associations that stick even when we know, in theory, they are not real.