In the weeks leading up to the June 2024 European Parliament elections, several far right political parties across France, Italy, and Belgium quietly made use of unlabelled artificial intelligence content to influence voters, despite public commitments to follow ethical standards in digital campaigning. The tactic allowed political messaging to spread rapidly online while avoiding transparency about how the material was created.
Investigations carried out DFRLab, Alliance4Europe, and AI Forensics uncovered at least 131 pieces of content that were either fully generated AI or significantly manipulated using AI tools. None of this material was clearly labelled as synthetic. The content circulated widely on major platforms including Instagram, Facebook, X, Telegram, and Vkontakte during the critical final weeks before voters went to the polls.
Researchers found that the material took many forms. Some posts relied on entirely fabricated images designed to appear realistic, while others used AI enhanced visuals to exaggerate emotions or dramatize political messages. There were also numerous examples of shallowfakes and cheapfakes, which are low effort manipulations that combine misleading captions with out of context or altered images. While technically simple, these tactics proved effective in capturing attention and spreading quickly through algorithm driven feeds.
According to analysts involved in the investigations, the goal of the content was not always to persuade directly, but rather to deepen existing social divisions. Many posts pushed conspiratorial narratives, amplified fear around migration and security, or portrayed political opponents in distorted and emotionally charged ways. blurring the line between real and artificial content, the campaigns contributed to a more confused and polarized online environment.
What has raised particular concern among researchers and democracy advocates is that some of the same political actors who benefited from these tactics are now involved in shaping future rules around AI and digital campaigning. This creates a troubling dynamic in which those who exploited regulatory gaps may gain influence over how those gaps are later addressed.
Experts warn that the lack of clear labelling allowed AI generated content to blend seamlessly into everyday political discourse, making it difficult for users to distinguish between authentic material and manipulated messaging. This undermines informed decision making and weakens trust in democratic processes, especially during elections.
The findings have renewed calls for stronger enforcement of transparency requirements and clearer rules around political use of AI. While many platforms have policies requiring disclosure of synthetic content, researchers say enforcement remains inconsistent and often too slow to respond during fast moving election cycles.
As generative AI tools become cheaper and easier to use, investigators warn that similar tactics are likely to appear in future elections unless safeguards are strengthened. The 2024 European Parliament vote has become an early case study in how unlabelled AI can quietly shape political outcomes, raising urgent questions about accountability, regulation, and the future of digital democracy in Europe.
