The “entry” of their political life with what this may entail for the voters’ choices, is documented by a Politico article and gives as an example the two confrontations of the last few days, in and in .
Connolly denounces an attempt to mislead voters
More specifically, according to what Politico reports, in Ireland a few days before the presidential elections, a deepfake video was released on the internet which showed the eventual winner of the elections, Catherine Connolly, supposedly withdrawing from the race. So according to testimonies many voters when they got the ballots were surprised by the fact that her name was on them.
The fake video with Connolly was described by her as a “shameful attempt to mislead voters and undermine our democracy”.
It featured a fake bulletin from Ireland’s national broadcaster RTÉ, which showed a deepfake version of Connolly saying: “It is with great regret that I announce the withdrawal of my candidacy and the end of my election campaign.” But the video went further as it featured deepfake versions of two well-known TV presenters discussing the news and its impact.
Both Meta and YouTube removed the Connolly video from their platforms without specifying how long it had been online.
Apology from Wilders to Timmermans – The far right is champion in deepfakes
Similarly in the Netherlands, during the election campaign a Facebook page frequently uploaded fake videos of the head of the ballot of the Green Left-Labour (GroenLinks/PvdA) electoral alliance, Frans Timmermans.
Behind these incendiary deepfake posts – one of which showed him being arrested in handcuffs and another of him holding wads of banknotes – it was eventually revealed that two MPs from Geert Wilders’ far-right Freedom Party (PVV) were behind them. According to Politico, this revelation forced Wilders to publicly apologize to Timmermans.
The posts of far-right MPs, however, were not the exception. A study of around 20,000 posts about the Dutch election by researchers from the University of Amsterdam and the University of Mainz found that more than 400 of them were generated by Artificial Intelligence.
Far-right leader Geert Wilders’ party, the Party for Freedom (PVV), emerged as the first to use Artificial Intelligence. Over a quarter of the AI-based posts (120 in total) were traced to PVV-linked accounts.
Wilders himself launched the PVV’s campaign with an AI-generated video depicting a fictional future Netherlands living under Sharia law. Dutch weekly De Groene Amsterdammer reported that the video was created with OpenAI’s Sora.
Are the efforts of Brussels enough?
As Politico emphasizes in its related publication, the regulatory authorities in Brussels have set as their main priorities the integrity of the elections, the dangers of Artificial Intelligence and online disinformation, but precisely because of the piecemeal nature of the measures they have taken, they have ended up just observers in these elections.
As the technology to create AI content and the platforms to distribute it are mainly based in the US, all eyes are on Brussels for an EU-level response.
The EU Digital Services Act recognizes that platforms have a partial responsibility for addressing risks in electoral processes, and both Meta and Google have recognized that AI is a significant risk factor. But as the publication emphasizes, they should do more to ensure that this content does not affect the elections.
