Entirely homegrown or helped? Brussels starts asking about France’s riots

EU regulators examine foreign disinformation around the riots in France

Protests against the French government turned violent over the weekend, with looting in Paris and Marseille, and unrest spread to Italy and Belgium. By Monday the European Commission was meeting French authorities, and by Tuesday the digital enforcement authorities of all member states, to examine a second layer of the story – whether foreign campaigns amplified the violence online, and whether platforms let them.

According to the Financial Times, pro-Russian and far-right accounts linked to earlier disinformation campaigns spread violent videos to deepen the chaos. One X account with 5,000 followers posted a clip of protesters throwing fireworks that gathered almost 6 million views; the footage was recorded at protests in Tbilisi, Georgia, in 2024. Italy’s defence minister Guido Crosetto blamed the unrest on “a hybrid war that someone is waging on our nations” through social media.

David Colon, a Sciences Po researcher who studies disinformation, told the FT that influencers linked to the Russian operations Storm-1516 and Matryoshka amplified the events and spread false narratives, but did not start the protests and had no effect on their course. European intelligence officials say they have found no concrete evidence that Russian agents stoked the street violence directly, and that the simplest explanation remains homegrown radicalism. Experts also see signs of Iranian and Algerian activity, which the accounts named deny. France is the second most targeted country in continental Europe after Ukraine, according to the EU’s diplomatic service.

The Commission says national authorities are handling the situation for now, and that Brussels will step in if platforms fail to cooperate and the risk becomes systemic.

NATO’s StratCom centre ran its first crisis exercise against an AI-simulated public

For three days in Riga, from October 12 to 14, teams from 13 countries sat at parallel tables in an events hall and tried to steer a fictional country called Avenia through an information attack by its neighbour Tessare. The exercise, Synesis 2026, was the first of its kind by the NATO Strategic Communications Centre of Excellence, the alliance’s Riga-based centre for countering influence operations.

Each team worked against 500 “synthetic citizens” – AI-simulated people with political leanings, moods and regional attitudes, built with the University of Genoa on survey data checked by social scientists. It was the first exercise in which the audience was simulated by AI rather than played by people. The citizens answered polls and followed or unfollowed media outlets depending on how much they trusted them.

Centre director Jānis Sārts said the scenario was built on Russian doctrine and escalated to just below the threshold of NATO’s Article 5, the alliance’s mutual-defence clause. Teams that assumed their public trusted them did worse. Teams that used AI tools did better, until they started relying on them too much. The organisers also entered a fourteenth team run entirely by AI, and it stayed among the top five for the whole exercise.

The organisers stress that a simulation cannot replace evidence from the real world. One threat they did not model this time is disinformation aimed at AI chatbots, and they expect it to matter more as more people get their news from chatbots. The exercise will be repeated.

A Russian bot network fakes both the videos and their view counts before the US midterms

AI-fabricated clips of actors and musicians – Robert De Niro, Julia Roberts, Cher – attacking Democrats began appearing on X in early September. Bloomberg reviewed 130 videos and doctored newspaper front pages published between September 10 and October 9 and, with researchers, attributed them to Matryoshka, a Russian operation that floods platforms with clips impersonating established media. The fakes carry ABC, CBS or CNN logos and allege corruption by Democratic candidates; the campaigns named have denied the claims. In the races Bloomberg checked, the operation reached 29 of the 36 Democrats running for governor and 20 of the 35 running for Senate.

Snapshots archived by Antibot4Navalny, an anonymous research group that has tracked Russian bot activity since 2018, show view counts jumping as high as 200,000 within an hour of posting and then flatlining – the opposite of how organic views grow. One post with fake clips about three Democratic Senate candidates logged 50,839 views and a single repost in its first 20 minutes, then fewer than 700 views over the next three hours. “Views don’t normally drop off a cliff,” Darren Linvill, a Clemson University professor who studies media manipulation, told Bloomberg. The operators publish five to ten new stories every weekday, mostly between 9am and noon, and take weekends off.

X suspended 80 percent of the accounts Bloomberg saw promoting the midterm clips, and new ones keep replacing them. On October 14 the platform said it had suspended more than 13,000 accounts tied to suspected influence operations linked to Iran, Russia and China. On October 20, four senior Democrats on the congressional intelligence and election committees wrote to 25 technology companies, including Alphabet, Meta, OpenAI and TikTok, asking them to remove false voting information faster, give election officials dedicated reporting channels and open their safeguards to independent researchers. Federal agencies say their election security support continues.

Taiwan’s security bureau counts a 48% rise in Chinese disinformation before November’s local elections

Taiwan’s National Security Bureau, the island’s main intelligence agency, told lawmakers it recorded more than 2.5 million pieces of Chinese disinformation in the first nine months of 2026 – up about 48 percent from 1.69 million in the same period a year earlier. The bureau connects the rise to the local elections Taiwan holds in November.

The content targets energy policy, food security, defence and individual candidates. The bureau says China produces it with AI tools, content farms and public relations firms – a setup that makes each item cheap and fast to turn out, so volume replaces craft. Beijing has previously denied interfering in Taiwan’s elections.