Elections everywhere are increasingly affected by disinformation campaigns, and this year's French municipal elections are no exception, with watchdogs identifying fake campaign websites. However, their local nature may offer them some protection from the effects.
France’s local elections are, in one aspect, particularly vulnerable to manipulation, with the large number of candidates making it easier for false information to circulate unnoticed.
“We are traditionally particularly attentive to the local elections, because it multiplies the possibilities of identity theft,” explains Pierre Bonis, head of Afnic, the organisation which operates France's country code domain names, including ".fr".
Election manipulators often try to make fake candidate sites using a technique called “typosquatting” – or URL hijacking – which diverts users to sites designed to look like those of candidates or local news sites, but containing false information.
While the large number of candidates – there were more than 900,000 candidates on more than 50,000 lists in the first round – exposes the French local elections to such risks, the nature of society in small villages and towns can limit the impact of disinformation – according to Bruno Courtois, a cyber security adviser at the Sopra Steria digital consultancy.
“Local elections are still based on trust and knowledge of local teams, which remains the best defence against disinformation,” he told RFI. Even in towns of up to 100,000 residents, elected officials and residents often know each other, which makes deception more difficult.
French cyber experts reveal vast network of Russian disinformation sites
Larger cities at risk
However, lager cities may be more vulnerable.
“What is more worrying is large metropolitan areas where the relationship between mayors and voters is more distant,” said Courtois. “I worry more about big cities than small and medium-sized ones.”
Afnic identified almost 200 fake websites during the municipal election campaign, Bonis told RFI. Many were set up well in advance of the election to build credibility over time.
One target was Pierre-Yves Bournazel, the candidate from President Emmanuel Macron’s Horizons group for mayor of Paris.
According to Viginum, which tracks foreign disinformation campaigns, an attack was carried out by the Storm 15-16 network – which is linked to Russian disinformation players that have been active in France and other European countries.
Anti-disinformation tools
Wikipedia, the collaborative online encyclopaedia – which can be edited by members of the public – sees a flurry of activity on candidate and issue pages around elections.
Wikimedia France worked with a data analysis company to create a barometer to note which pages are most often modified ahead of elections, in order to inform moderators, journalists, researchers and the general public.
But it is impossible to track all manipulation on Wikipedia or elsewhere, and simply shutting down fraudulent sites is not enough to address the issue.
“It goes beyond elections. It is about protect the integrity of the European information space more broadly,” said Courtois, pointing to initiatives such as media certification systems, or using tools to label content generated by artificial intelligence (AI) on social media platforms.
Another priority is strengthening media comprehension and social awareness through education.
“There are all the education efforts that can be put in place on these issues, but the reality is that it is difficult and takes time,” he said.
‘A new battleground’: France takes its fight against disinformation online
AI enters the game
Disinformation around elections is not new. Courtois points to the Ems Dispatch in 1870, a telegram edited by Prussian Chancellor Otto von Bismarck to provoke France into declaring war on Prussia, as an early example of the consequences of disinformation.
What has changed, he says, is the speed and reach of information through digital technology.
“It’s this ability to reach any citizen that makes this threat more widespread, personalised and harder to contain."
AI too has become part of the electoral landscape, used by campaign teams and voters alike.
“It’s a tool that has become totally essential,” said Clarisse de Maupeou, an AI specialist at polling agency Ipsos, which published a survey in which 25 percent of respondents said they used a generative AI model to learn about a politician or party, and 23 percent said they would consider doing so.
“As for how people use it, they do not ask outright who to vote for. They ask it to help them identify which candidate best matches their views.”
She notes that people have always used available tools to inform themselves.
”In the past it was newspapers and television, then social media. What is most practical for young people today is using AI,” she said.
She raises concerns, however, about the origins and potential bias of AI tools – many of which come from the United States.
For Bonis, the election itself is more of a target than any individual candidate.
"The ultimate aim of disinformation during elections is to put people off voting. It is not necessarily about convincing them to vote for one candidate or another. The more confusing the information, the more it creates a [general] sense of distrust.”
