Importance Score: 75 / 100 🔴
Italian Opposition Condemns Alleged Racist AI Images by Deputy PM’s Party
Rome, Italy – Opposition parties in Italy have formally complained to the national communications watchdog regarding a series of AI-generated images disseminated on social media by Deputy Prime Minister Matteo Salvini’s far-right League party. The images are being denounced as “racist, Islamophobic, and xenophobic.” The controversy highlights growing concerns over the use of artificial intelligence in political propaganda and the spread of potentially harmful content online.
Formal Complaint Lodged with Communications Authority
The center-left Democratic party (PD), in conjunction with the Greens and Left Alliance, submitted a formal grievance to Agcom, Italy’s regulatory body for communications. The complaint alleges that the contentious images utilized by the League party embody “almost all categories of hate speech”.
Controversial AI-Generated Content Disseminated Online
Over the preceding month, numerous images, purportedly created using artificial intelligence, have surfaced on the League’s social media platforms, including Facebook, Instagram, and X. These images frequently depict people of color, often wielding knives, and appearing to assault women or law enforcement officers. Critics argue these depictions contribute to xenophobia and racial prejudice.
Senator Accuses Party of Inciting Hatred and Deception
Antonio Nicita, a senator from the PD, asserted, “The images published by Salvini’s party and generated by AI embody nearly every type of hate speech, spanning from racism and xenophobia to Islamophobia. They are weaponizing AI to target specific groups – immigrants, Arabs – who are portrayed as potential criminals, thieves, and rapists.”
Nicita further elaborated, “These images are not only violent but also misleading. By obscuring the faces of alleged victims, they create a false impression of authenticity, deceiving users into believing the photos are genuine. These are images designed to stoke animosity and incite hatred.”
vCard.red is a free platform for creating a mobile-friendly digital business cards. You can easily create a vCard and generate a QR code for it, allowing others to scan and save your contact details instantly.
The platform allows you to display contact information, social media links, services, and products all in one shareable link. Optional features include appointment scheduling, WhatsApp-based storefronts, media galleries, and custom design options.
MP Highlights Strategy of Fear-Mongering
Francesco Emilio Borrelli, a Member of Parliament representing the Greens and Left Alliance, echoed these concerns, stating, “This situation is grave. AI generates content based on our instructions, and in this instance, it was clearly instructed to produce images of Black individuals robbing elderly or frightened women. This is part of their deliberate strategy to cultivate fear among the populace.”
League Party Confirms Digital Image Generation
A spokesperson for Salvini’s party acknowledged that “some of the pictures” disseminated on their social media channels were indeed “generated digitally,” confirming the use of AI image generation technology.
Party Defends Content, Cites Real-World Events
In an official statement, the League party defended its actions, asserting, “The core issue is not the image itself, but the underlying reality. Each post is rooted in factual reports from Italian newspapers, complete with names, dates, and locations. If reality appears harsh, the blame should fall not on those who report it, but on those who create such a reality. When reporting on crime, cheerful or reassuring imagery is simply inappropriate.”
AI Forensics Expert Raises Concerns Over Realism
Salvatore Romano, head of research at AI Forensics, a non-profit organization, confirmed that the League’s images exhibited “all the characteristics of artificial intelligence“. He pointed out, “They are decontextualized photos where the subject is prominently in the foreground while the background is completely blurred. My concern is that these AI-generated images are becoming increasingly realistic,” making it harder to distinguish them from genuine photographs.
Complaint Cites Misleading Use of Media Branding
The formal complaint submitted to Agcom provided several examples of images believed to be digitally fabricated. It highlighted that these images were often presented alongside the branding of established mainstream media outlets. While these outlets had reported on the crimes referenced, they had not utilized images of the alleged perpetrators in their actual reporting, making the League’s use of AI-generated content appear misleading.
Specific Examples of Contentious Posts
Case 1: Train Conductor Attack
One League post featured the text: “A foreigner attacks the train conductor,” accompanied by an AI-generated image of a man of color with a raised fist. The original news headline from Il Resto del Carlino read: “He attacks the [female] train conductor and sparks panic on board.” The news article referred to the suspect as a “foreigner” but provided no further details regarding nationality, and notably, no photograph of the alleged incident was published.
Case 2: Family in Islamic Dress
Another image cited in the complaint depicted a mother and father in Islamic attire seemingly yelling angrily at a young girl. Critics argue this image promotes “racial and Islamophobic prejudice.” Il Giorno, the newspaper cited by the League, reported on the case but made no mention of the family’s or the girl’s religion, only noting that the child attended an Arabic language school. The original report also did not include any photographs of the family.
Growing Trend of AI Propaganda by Far-Right Groups
The utilization of AI-generated images for propaganda, particularly by far-right political parties, is an escalating phenomenon. This trend gained prominence around the time of last year’s European elections, when images designed to amplify fears about immigration and denigrate political figures like Emmanuel Macron began circulating widely on social media platforms. Concerns regarding political deepfakes and manipulated media are on the rise.
Romano added, “The American elections, with figures like Donald Trump and Elon Musk, effectively normalized this trend. Today, we observe that far-right parties have not only continued to generate fake images for propaganda purposes but have amplified their use. This is occurring as AI tools improve content quality, making the phenomenon increasingly worrisome and harder to counteract.”
Social Media Platforms and Content Moderation
Despite obligations for social media platforms to mitigate these risks, such as by labeling AI-generated content, Romano suggests that these measures are often ineffective. The rapid advancement and sophistication of AI image generation pose significant challenges for content moderation and detection of misinformation.
League Party Responds to Hate Speech Concerns
When questioned about whether the League recognized that the images could incite hate speech, a party spokesperson responded, “We regret this situation, but our sympathy lies with the victims, not the perpetrators. If denouncing crimes committed by foreigners is labeled ‘xenophobia,’ perhaps the issue lies not with the term, but with those who use it to stifle debate. We will continue to denounce, with strong language and images, what others prefer to ignore.”
Potential Regulatory Consequences
If Agcom determines that the flagged content is indeed offensive, it possesses the authority, under the EU’s Digital Services Act, to mandate the removal of posts, suspend accounts, and impose fines on social media platforms for inadequate policing of user conduct. In 2023, Agcom fined Meta €5.85 million and ordered the removal of numerous accounts for violating regulations against gambling advertising, illustrating its power to enforce online content standards.
Social Media Platform Responses
Meta, the parent company of Facebook and Instagram, was contacted for comment but did not provide an immediate response. A spokesperson for X stated, “We are not legally obligated to label every AI-generated image. It is quite evident that these posts are straightforward political campaigning.”
The X spokesperson added, “Rest assured, we are committed to maintaining the authenticity of discussions on X, and we ensure we are well-equipped to combat manipulated media, including the increasing trend of ‘deepfakes’. We also apply visible labels to any such content that credible sources have debunked.”