A Russian propaganda operation targets France with AI-generated fake scandals, drawing 55 million views on social media.

Questo articolo è disponibile anche in: Italiano

A Russian propaganda operation that previously targeted elections in the United States and Germany has now set its sights on France. NewsGuard has identified that the Russian influence campaign known as Storm-1516 struck France with five false narratives circulated between December 2024 and March 2025. John Mark Dougan, a former deputy sheriff from Florida who was granted asylum in Moscow and became a Kremlin propagandist, plays an active role in Storm-1516. He uses artificial intelligence to create and disseminate false claims across his network. These disinformation narratives targeting France appeared in 38,877 social media posts, generating 55.8 million views. For comparison, in the previous four months, only one fake news story from Storm-1516 targeted France, spreading across just 938 posts and garnering 845,000 views. Additionally, NewsGuard analysts found that major generative AI chatbots have echoed these false narratives about France, highlighting a new threat in the disinformation landscape. False claims are now reaching the public not only through social media but also via tools people use to search for and consume news and information. A retaliation for Macron’s support of Ukraine These disinformation campaigns emerged as French President Emmanuel Macron increased his military support for Ukraine, while U.S. President Donald Trump retreated from involvement. At the same time, France is facing domestic political turmoil, with far-right leader Marine Le Pen—supported by Russia—being disqualified from the 2027 presidential race following a conviction for embezzlement, pending appeal. Since late last year, NewsGuard has identified five viral false claims targeting France:

• A video allegedly showing a migrant from Chad confessing to the rape of a 12-year-old girl in France;
• A false report claiming Ukrainian President Volodymyr Zelensky secretly purchased the French private bank Milleis Banque for €1.2 billion, part of a Russian disinformation effort to accuse him of misusing military aid for personal gain;
• A video seemingly depicting members of the Islamist militant group Hay’at Tahrir al-Sham threatening to burn down Notre-Dame Cathedral;
• A video in which a French AIDS activist from the Caribbean, named Réaulf Fleming, claims his deceased brother had a relationship with Emmanuel Macron;
• A video of an alleged former student of French First Lady Brigitte Macron accusing her of sexually abusing him when he was 12 years old.

A European security source confirmed to NewsGuard that all five narratives originate from Storm-1516, a spinoff of the Russian troll farm Internet Research Agency. “We are detecting massive interference from Russia” French officials have acknowledged that Russian disinformation attacks are increasing. “After Ukraine, France is the European country most targeted by foreign manipulation attempts,” said French Prime Minister François Bayrou at a forum organized by the French government agency Viginum on March 28, 2025, according to Agence France-Presse. “We are detecting massive interference from Russia,” he added, emphasizing that artificial intelligence “enables these manipulation campaigns to be carried out on an unprecedented scale.”

The Russian disinformation operation Storm-1516 continues to pose a significant and growing threat to global public information. According to NewsGuard, among the five false claims that targeted France between December 2024 and March 2025, the one with the highest engagement alone reached 17 million views, a figure obtained using advanced media analysis tools.

Since 2023, NewsGuard has identified 51 false narratives linked to Storm-1516, aimed at strategic targets such as Ukraine, the 2024 U.S. elections, the 2024 Paris Olympics, the 2025 German elections, and now France. Despite the U.S. sanctions on Valery Korovin, considered one of the main architects of the operation, Storm-1516 continues to spread disinformation globally, seemingly without encountering significant obstacles.

The combination of advancing generative artificial intelligence models and the reduction of content moderation efforts by platforms has allowed these campaigns to gain speed, credibility, and penetration. False narratives not only invade social networks but are also assimilated and repeated by AI chatbots, making it harder to distinguish truth from falsehood.

The five false claims targeting France circulated in nearly 39,000 posts online, 1,284 of which were from accounts located in France, and spread across over ten platforms — including X, Facebook, Reddit, Instagram, Rumble, and Bitchute — in a dozen different languages, from French to Albanian, and even Vietnamese.

A NewsGuard audit revealed that the most widely used Western chatbots repeated Storm-1516’s false narratives 32% of the time. This is made possible by a strategy of “recycling narratives,” where fake content is spread through fake local news websites and YouTube videos, only to be recognized as “reliable sources” by chatbots.

An emblematic example is the false story about a supposed relationship between President Emmanuel Macron and a man who died of AIDS. The March 28, 2025 YouTube video that originated the narrative features a supposed Caribbean activist, Réaulf Fleming, claiming that his brother (actually his cousin) had been involved with Macron. When NewsGuard asked 11 chatbots to comment, 5 repeated the false claim as fact, including Le Chat, the chatbot developed by the French company Mistral.

Despite the lack of verifiable evidence, Le Chat claimed that “Réaulf Fleming said that his brother Dimitri had a romantic relationship with President Macron and then died of AIDS,” a response that highlights the vulnerability of current defenses against disinformation integrated into AI systems. NewsGuard requested clarification from Mistral in April 2025 but received no response.

Mistral repeats a false claim about Macron. (Screenshot by NewsGuard)

Six out of the eleven chatbots rejected the claim. For example, You.com responded: “There is no proof or mention that Réaulf Fleming claimed his brother had a romantic relationship with French President Emmanuel Macron and died of AIDS.”

The chatbots behaved inconsistently when asked about other false narratives related to France.

In addition to being repeated by some chatbots, the false story went viral on social networks, following a pattern of circulation that has become common. One day after the YouTube video was uploaded, the claim was picked up by SeneNews and ActuCameroun, two West African news websites that, according to NewsGuard’s analysis, had previously published Russian disinformation and had been cited by some chatbots in their responses.

On April 1, 2025, the false claim was spread by the Pravda network, consisting of around 150 pro-Kremlin sites based in Moscow and operated anonymously. These sites recycle disinformation in multiple languages, seemingly to contaminate web crawlers and generative AI. The claim generated 14.6 million views.

A false narrative targeting Macron spread from an anonymous YouTube channel to pro-Kremlin media.

Eye for an Eye

A French military officer involved in cybersecurity, who requested to remain anonymous for security reasons, told NewsGuard in a phone interview in April 2025 that “Russian disinformation operations targeting France are intensifying as France’s support for Ukraine strengthens and as France takes full responsibility for it.”

False claims targeting France emerged in conjunction with key political events. According to NewsGuard‘s analysis, the false narratives targeting France appeared in parallel with France’s demonstrations of support for Ukraine, such as diplomatic visits, official statements from Macron, and high-profile meetings. This deliberate timing became a hallmark of the Storm-1516 project: NewsGuard had previously identified that false corruption allegations against Zelensky emerged in response to his travels, accusing him of purchasing properties in countries he visited. The exploitation of a public employee’s identity in a defamatory campaign against Brigitte Macron Storm-1516 has refined a now well-known tactic: selecting a real person with a verifiable link to a public figure and constructing a false accusation about that figure, using AI, fake social media accounts, and real biographical details drawn from yearbooks and alumni networks to give the narrative a veneer of authenticity. In early February 2025, a video uploaded on X showed a man named Lionel Torres accusing Brigitte Macron of sexually assaulting him when she was working as a teacher at a middle school in Strasbourg, in eastern France. “When I was 12, I was sexually abused by someone who is now very famous in France,” said the man presumed to be Torres in the video. Mixing real yearbook images with an AI-generated video impersonating Torres, the alleged Torres describes a gruesome but entirely fabricated story. The video is false: confirmation from the alleged victim The real Torres, a 50-year-old public employee living near Lyon, told NewsGuard in a video interview in April 2025 that he was indeed a former student of Brigitte Macron, but he had never recorded such a video and had never made any such accusations. “They must have taken my picture, I suppose, and reproduced it on a 3D model or something like that,” he said. “I am a victim in this case, we can agree on that… And when I say victim, I’m talking about the [false] video, of course.” It appears that face-swapping technology was used to overlay Torres‘ facial features onto someone else’s. The video shows signs of manipulation done with AI, including unnatural and inconsistent movements, audio out of sync with the lip movements, and a strong Slavic accent rather than a French one.

Real footage and AI tools were used to exploit Lionel Torres’ identity and create a disinformation campaign. (Screenshot via NewsGuard) Tactics of recycling without borders The case of Torres echoes a disinformation narrative that went viral during the 2024 U.S. presidential election, when a fake video created using artificial intelligence appeared to show a former student named Matthew Metro accusing then Democratic vice-presidential candidate Tim Walz of sexually assaulting him while Walz was a teacher in Minnesota. Metro told the Washington Post he had never met Walz. Similarly, a video from December 2024 showed an 18-year-old named Milina Graz falsely accusing the German Green Party chancellor candidate (later defeated) Robert Habeck of sexually abusing her in 2017. What now? So far, France has been spared at least one tactic typically used by Storm-1516: the use of a coordinated network of websites posing as local news outlets. In both the U.S. and Germany, AI-generated hoaxes were subsequently spread by websites pretending to be local news sources, created and managed by Dougan. His networks of fake local news websites were pivotal in circulating Storm-1516 content, giving false narratives more credibility and helping them spread. 171 fake local sites were created in the U.S. ahead of the 2024 election, and 102 before the German election. A French official said France is preparing for this phase of the Russian disinformation operation. “Every time France’s support for Ukraine strengthens, Storm-1516 is one of the first groups to act. Others follow suit,” the French military official said to NewsGuard. “It’s like with cyberattacks: we observe the tactics, techniques, and procedures, and this allows us to be aware that probably, other strategies of the Russian disinformation machine will be applied, and we are preparing for that.” A cura di Natalie Huet, McKenzie Sadeghi e Chine Labbe Supervisionato da Dina Contini ed Eric Effron