Why far-right parties are using artificial intelligence in election campaigns

Images of Muslims chased by giant pigs, pregnant men, plates overflowing with insects, burning flags, immigrants eating dogs and cats. THE far-right parties they took AI models and turned them into AI machines mass disinformationto fuel hatred, fear and prejudice. According to experts, the 2024 European Parliamentary elections were a test case electoral campaigns based on artificial intelligence.

Not only in Europe, looking at more moderate groups, the Republican Party also churned out fake images during the US elections to give visual evidence to conspiracy theories and push voters to vote for Trump. “What we’re seeing it’s just the tip of the icebergbecause content generated by individuals and outside official channels is even worse,” he explained to the British newspaper The Guardian Salvatore Romanohead of research at the organization non-profit AI Forensics.

In fact, the goal of the images is not so much to deceive people. They work more like memes, they want provokeeven entertain, and go to reinforce electoral rhetoric who doesn’t need real facts to destroy the enemy. “On the one hand these images can influence elections or compromise emergency responses in crisis situations, on the other hand they start a process of erosion of trust towards him information bodiesespecially when these fake news are personalized”, explained Giulio Corsi, researcher at the Institute for Technology and Humanity of the University of Cambridge, to Fanpage.it.

How AI models are used for propaganda

Software excels at producing hyperrealistic content because they have been trained on tens of thousands of photographs and drawings published on the web. Anyone can log in and create with a string of words fake images that look real. As Corsi explained: “Artificial intelligence models have introduced new complexities and risks into information environments, synthetic media can in fact facilitate the spread of false news and erode public trust.”

Not only that, second William Allchornsenior research fellow at Anglia Ruskin University, “artificial intelligence lowers the barriers to entry for content creation. No programming skills required or things like that to generate these images, and so the far-right opinions are becoming mainstream or normalized,” he explained. The far right has not raised moral issues on images created with artificial intelligence for propaganda, “the more consolidated parties”, however, “still have great fears about the ethics and reliability of these models”, explained Allchorn.

The images created by far-right parties

There are images of women and children eating insects, Islamophobic memes, police officers trampling on national flags. They all fuel feelings of fear by going to radicalize already extreme positions, which often lead to racism. According to experts, in almost all cases far-right parties and movements have avoided using any type of watermark on images created by artificial intelligence.

The far-right German party Alternatives for Germany (AfD) built its election campaign on images created with AI. He posted a drawing of a blonde woman with blue eyes raising her hand in a stop sign with a German flag in the background. Next to the image we read: “People must stop coming to Germany because the country no longer needs them gang rape specialists“.

In September, the Brandenburg branch of the AfD published similar images and pro-AfD groups on Facebook, in addition to resharing images from the election campaign, also created an image of a giant pig chasing a group of people in Islamic dress, accompanied by the slogan “Arabic film version of Godzilla“. The images created with AI also arrived on the profile of Matteo Salviniwho shared both a fake photo of a pregnant man and a post showing Muslims burning the divine comedy.