A mass of AI-generated videos of protests across Europe has emerged on TikTok, which experts say aims to "amplify societal polarisation".
A series of AI-generated videos showing thousands of people protesting in a range of European cities against immigration, the state of the country, the current government, or for nationalist values has been doing the rounds on TikTok.
ADVERTISEMENT
ADVERTISEMENT
The videos analysed were published in the past four months, with some gathering more than 300,000 views.
The Cube, Euronews' fact-checking team, has identified this kind of content targeting users in Portugal, France, Germany, Ireland, Italy, and the UK, with many appearing to believe what the videos showed.
However, none of the protests depicted in these specific videos is real, yet they use tactics to try and fool users into thinking otherwise.
To begin with, many of the videos pose as legitimate news outlets.
For example, one of the videos appears to show a Portuguese journalist working for the state-owned media RTP, as seen by the logo on her microphone, while the graphics embedded into the video itself resemble those of the Portuguese private media channel SIC Notícias.
Her location is also off: in real life, the Tagus River flows where the building she is standing in is depicted.
The Cube also reached out to RTP, which told us that the journalist in the AI-generated video did not work for them, and that they have no connection whatsoever with the video in question shared on TikTok.
Despite the vague reasons behind these alleged protests, such as those expressing anger at "the state of the country", some messages in the videos have connections with far-right parties. This is the case in some German videos that The Cube team has encountered, which contain captions in support of the Alternative for Germany (AfD) party.
Some also contain comments from TikTok users who evoke these parties, asking them to step up and fix issues such as migration. Some of the figures called upon include André Ventura, of the Portuguese far-right party Chega; the French National Rally; and Germany's Alice Weidel, from the AfD.
Most of the videos end by asking the audience to engage with their content by sharing their opinions about the protests or the name of the city they are watching from.
Experts claim that the primary purpose of the mass production of these types of videos is to manipulate public perception and amplify societal polarisation across Europe, posing severe risks to democratic discourse.
"Audiences engage with these videos primarily because synthetic imagery taps directly into pre-existing political anxieties and cultural grievances," Bruna Martins dos Santos from WITNESS, a global organisation that examines the impact of emerging technologies, especially generative AI and deepfakes, told The Cube.
Familiar misinformation tactics
All of these videos use three common misinformation techniques: emotional language, AI-generated content and polarisation.
Most of these tactics serve as engagement bait, according to Martins dos Santos.
"Topics like immigration, nationalism, and societal decline evoke strong emotional responses," she said. "This effect is especially pronounced on short-form video platforms where visual impact and emotional resonance precede context, leading viewers to process the spectacle — the crowd size, flags, and chanting — long before questioning its authenticity."
Additionally, social media recommendation algorithms amplify this dynamic by pushing provocative visual content to users most likely to resonate with it.
Using AI to produce this content makes the production cheap and scalable, effectively driving views, follower counts, and interaction metrics, Martins dos Santos said.
Why is this kind of content available?
The videos have been labelled as AI by TikTok itself. The platform claims that it bans AI-generated content that is "misleading about matters of public importance" or violates its misinformation guidelines.
In fact, it goes further: "Unlabeled content may be removed, restricted, or labelled by our team, depending on the harm it could cause."
Some videos, however, seem to slip through.
"The fact that these videos remain online can mean a policy violation of TikTok's own rules, but also highlight a failure in the platform's detection tools' capacity," said Martins dos Santos.
Moreover, under the European Digital Services Act, it must also take effective measures against content that could negatively affect "civic discourse and electoral processes".
"While the continued presence of these videos does not automatically establish a policy breach, it raises critical questions about the effectiveness of TikTok's detection mechanisms, labelling standards, and recommender algorithms," Santos said.
TikTok did not immediately respond to a request for comment.
Real protests have gripped Europe in recent weeks
While the AI-generated videos show fabricated protests that did not take place, there have been widespread demonstrations across Europe recently for a range of issues.
Perhaps the most high-profile has been the protests in Spain in response to the Ceuta crisis, which saw an estimated 80,000 people breach the border between Morocco and the north African Spanish territory.
The majority of the migrants have since returned, but estimates say that up to 13,000 of them could still be in Ceuta, prompting protests in response to the government's handling of the crisis.
In Germany meanwhile, demonstrations erupted in opposition to the AfD party after it won a landslide victory in a state election earlier in September.
Thousands in Switzerland, on the other hand, recently marched to protest government inaction in the face of global warming.