Logo

How AI Became the Far Right’s Newest Weapon Against Refugees

From deepfakes to election propaganda, European groups are using cutting-edge technology to smear refugees and entrench prejudice

Share
How AI Became the Far Right’s Newest Weapon Against Refugees
A supporter of the far-right Alternative for Germany (AfD) party waves a German flag at left-wing protesters in Dresden, Germany, in 2024. (Sean Gallup/Getty Images)

Earphones plugged in, a cigarette dangling from his hand, Mohammed al-Mohammed was waiting for a train at a station in Hamburg, Germany, when he was enveloped by screams of horror. Amid the chaos of people fleeing, he glimpsed a flash of metal out of the corner of his eye. Turning, he saw a lady clutching a raised knife, pointed toward him. 

Mohammed’s reflexes kicked in, and he pushed the lady away. As another man tackled her and threw her to the ground, Mohammed pinned the lady down until the police arrived. They congratulated him, treated him to a cappuccino and the matter was closed. The next day, as part of interviews he did with media outlets, a German tabloid took a picture of Mohammed at the station, in which he posed with his arms on his waist, grinning victoriously. 

A couple of days later, Mohammed’s friend messaged him. His photo at the station had been manipulated using generative artificial intelligence (AI), depicting him in front of fake backgrounds, used as fodder for sardonic comments about what was perceived as the glorification of Mohammed’s act. Other photos showed an AI-generated likeness of him, leading to internet users questioning the veracity of the event. The photos made the rounds online, helping to propagate a large-scale disinformation campaign in Germany. The far-right politician Marie-Thérèse Kaiser tweeted under a photo of him, offering a “ticket for sending him back to his country.” 

“The bad reaction honestly devastated me deeply, because my act was just human,” Mohammed said. “I didn’t do anything wrong, yet the racists were coming at me.” 

The civil war in Syria started when Mohammed was 5 years old. Now 19, he made it to Germany after an arduous journey, having tried multiple times to escape the war, with border police even throwing him into a river during one of his attempts. He arrived in Germany in September 2022 to seek asylum, but his application was rejected and he is currently in “Duldung” status, which means his deportation is temporarily suspended.

Germany is the world’s third-largest host of Syrian refugees, who were pushed out of their country by years of civil war and a shattered economy. Compounding the all-too-real issue of problematic media portrayals of Syrian refugees is the weaponization of ever-evolving AI by far-right parties like Kaiser’s Alternative for Germany (AfD). 

Ahmed al-Kurdi, a Syrian stand-up comedian based in Germany, was part of a team that made a movie in 2020 titled “In Meinen Augen” — German for “In My Eyes” — which explored the media’s portrayal of Syrians in Germany through the eyes of Syrian immigrants and refugees. Al-Kurdi explains how so much of the visual coverage of the refugee crisis was done in a way that created distance between German readers and Syrian migrants. For instance, cameras placed immigrants in front of a row of policemen, congregated in large groups behind gates, or “lower than the police,” with photos taken from a higher point of view. This issue is not specific to Germany. Despite largely sympathetic coverage of the Syrian refugee crisis in early 2015 throughout Europe, the angle changed to “suspicion, and in some cases, hostility towards refugees and migrants,” according to a 2017 Council of Europe report. The paper also explained how press coverage consistently promoted hate speech and anti-migrant sentiment in some parts of Eastern Europe.

Salvatore Romano and Natalia Stanusch, researchers at the European nonprofit organization AIForensics, say that AI chatbots draw upon previously published material to create a realistic figment. According to a 2024 report by the Kofi Annan Foundation, nearly every national election globally since mid-2023 has seen the use of forms of generative AI. 

In AIForensics’ research, they delve into the weaponization of generative AI, particularly during elections. One pattern they have uncovered is of a “preference in some far right parties or communities [toward] these types of images on social media … across different countries.” 

“We did not necessarily find images that would show completely made-up stories. … It was rather this hyperbole of the type of visuals that you would encounter in the media, just absolutely portrayed in a very dramatized and negative way,” Stanusch said. 

In a 2024 report, the team uncovered multiple AI-generated images used by the National Rally, a right-wing nationalist party in France, as part of their “Europe without them” campaign. Some of the images included AI-generated depictions of migrant-filled boats, with the caption “Stop massive migration.” Others were blatantly Islamophobic, with an AI-generated image of what appeared to be a Muslim scholar within the walls of a mosque, captioned “[President Emmanuel Macron and European Commission President Ursula von der Leyen] compromised themselves with Islamists.” 

“[The images] sort of took something that was somewhat of a visual reference and just pushed it to the edge,” Stanusch said. This is worsened by social media algorithms that tend to “push content that gets viral.”

A September 2024 study by the University of Potsdam, for instance, showed that the AfD experienced double the success of other parties among first-time voters on TikTok, who were, on average, exposed to more than one video per day that “thematizes AfD content,” like anti-immigration sentiment and policies. The study also showed that users were exposed to nine videos per week of AfD content, as opposed to about one video of more left-leaning parties. 

Right-wing rhetoric is “more prone to being exacerbated by these kinds of images,” according to Romano, who added that generative AI “is definitely pushing the narrative and adding something to it, and I really can’t see any positive effect of this.”

Al-Kurdi’s participation in the production of “In Meinen Augen” was partly driven by his own tale of horror from an attempt to flee Syria. When he first tried to escape in September 2013, his boat was stopped near what he believes was the Greek island of Samos. He laughs dryly as he recounts the story. “I still remember this motherfucker — a guy from the coast guard group — he said to me, ‘Good luck next time.’” 

At the closest point between the two, Samos is only 1 mile from Turkey. Bordering Syria’s rebel-controlled Idlib province at the time of Bashar al-Assad’s rule, Turkey was a popular first destination for many Syrians hoping to start a new life abroad. A report published in 2025 by I Have Rights and the Border Violence Monitoring Network (BVMN) illustrated how Samos today uses an “EU-funded, AI-led surveillance system.” Two systems employed on the island — Centaur and Hyperion — use AI to closely surveil those in the extremely tightly controlled Closed Control Access Center (CCAC) in Samos, where refugees are brought in, if not subjected to pushbacks, as in the case of al-Kurdi. 

According to the report, Centaur likely uses AI to collect and flag movements that are “potentially aggressive or hostile.” Hyperion, on the other hand, uses biometric data to regulate movement into and within the CCAC. This measure was limited to migrants; nongovernmental organization workers were not subjected to the same practice. 

Most residents in the CCAC were not aware of what their data was collected for. Such surveillance technology, specifically at a border which is an entry point for many Syrians, has an impact on human dignity, and “creates an‬‭ environment‬‭ of‬‭ discomfort‬‭ and‬‭ anxiety,” according to the report.

Nikolett Aszodi, previously a policy and advocacy manager at the NGO AlgorithmWatch, said that despite recent legislation like the EU Artificial Intelligence Act, which pushes for greater transparency in the use of AI in the public sector, migration is one of the fields in which no public transparency is required. “Systems in these areas will only be listed in a private section of an EU database, and we won’t have any public scrutiny,” she explained. 

In 2016, the EU brought the General Data Protection Regulation (GDPR) into effect to protect the privacy and security of those within its borders. In April 2024, Greece’s Ministry of Migration and Asylum was fined nearly 180,000 euros for breaching this law in multiple ways, including a lack of transparency and compliance in investigations. 

Certain technologies that employ AI in the field of migration, like lie and dialect detectors, could provide output that is used to justify the existing intent of lawmakers and stakeholders. Aszodi refers to the Systeem Risico Indicatie (SyRI) in the Netherlands, a welfare fraud detection system. 

“I encountered the issue of writing a code which was supposed to single out people with migrational background and hence flag them as ‘risky,’” she said. “It’s kind of like outsourcing a decision that was already made prior. It’s very sneaky.”

Stanusch said that a first step in solving the problems caused by generative AI on the internet would be for social media platforms to accurately label AI content as such. With algorithms often built to favor negative, heated content — as seen in the Potsdam University study — her second suggestion is more difficult to execute. 

In her view, we need “to come to terms with the fact that perhaps the business model that favours this type of recommendation system is not the best for our society, for us as humans, or for democratic processes. And pushing to recognize the fact that it’s the platform’s responsibility, the business model behind them, that is harming all of us. … It’s a very long-term process, but something that I think we should talk about way more.” 

The blitz overthrow of the Assad government by rebel forces in December 2024 brought an abrupt end to the country’s 14-year-long civil war. The fall of the regime sparked discussions in a number of European countries about the Syrian refugee issue. Less than 48 hours after Assad was toppled, Germany froze all its pending asylum applications for Syrians. Yet going back to their homeland is not often easy or viable for Syrians. With an economy in tatters and continued political instability, a return may imply closing the chapters of the life they have built abroad. 

Mohammed said that he is eager to continue learning German to work and support his family back in Syria, but is unable to do so without financial support. Despite multiple inquiries at the immigration office, he is constantly told to wait without a clear timeline. He is focused on improving his life in Germany, rather than considering a return to Syria. 

“Where do I return to? There’s no safety or future. Go back and die from hunger, to a house on the ground?” he asked. “I don’t want to go back there and lose myself.”


“Spotlight” is a newsletter about underreported cultural trends and news from around the world, emailed to subscribers twice a week. Sign up here.

Sign up to our newsletter

    Will be used in accordance with our Privacy Policy