WhatsApp campaign in Brazil

More than 10 million WhatsApp subscribers received instructions from Brazilian deputy André Janones on the day of Lula's re-election, which took place on October 25 with Lula securing 51.8% of the vote against Flávio Bolsonaro. Janones assigned daily "missions" to his channel "Exército do Lula" subscribers, who grew from about 50,000 before the first round to over 10 million.

French internet users adopt similar methods

French internet users have since adopted similar methods as the 2027 presidential election approaches, with some declaring in millions the blocking of far-right accounts, though the exact impact of such actions on algorithms remains unclear.

Student mobilization on social media

In France, students from the Antoine-de-Saint-Exupéry high school in Créteil organized a Snapchat group to block their school, and the movement spread nationwide. Students displayed their demands in Instagram "stories" and set up private loops on Instagram, Snapchat, or WhatsApp, sometimes with hundreds of members, according to Le Monde.

TikTok mobilization and protests

On October 17, TikToker V2 proposed a large mobilization on purchasing power, with content related to the demonstrations accumulating 800 million views on TikTok, according to Visibrain. The images of looting overshadowed the peaceful processions.

Communication during protests

For the second act on October 24, demonstrators exchanged free collaborative cards and instructions on Telegram, with some installing BitChat, a Bluetooth-based messaging app that allows maintaining contact during processions. The calls for the third act on October 31 circulated in the form of parodies, including an invitation to dance the Thriller choreography for Halloween, viewed nearly 1.3 million times on TikTok.

Engagement and social media impact

The real engagement is volatile and difficult to measure. A share or a view does not engage the author to the same degree as a march in a procession, and no social network publishes the number of spectators who go to gatherings after watching the call video. Sociologist Raphaël Lupovici noted that social networks do not mechanically dictate the vote, and their effects on voters, when measured, are minimal.

Blocking far-right accounts on X

French internet users have attempted to reduce the visibility of political accounts they oppose. On October 27, streamer Arkunir, followed by 1.1 million people on X, asked his subscribers to block every far-right account they encountered, a message viewed more than 3.5 million times. Websites such as Un Facho KO and Mise au Poing allowed users to block far-right accounts, with Un Facho KO reporting over 15 million declared blocks on October 30, and Mise au Poing reporting 3,083,584 on October 31.

Clarification and division in France

On October 28, André Janones denounced a "translation error," clarifying that he never advocated blocking far-right accounts but rather ignoring their publications. This clarification divided French participants, with La France insoumise no longer recommending systematic blocking, while Mise au Poing maintains its instruction. Antoine Léaument, a LFI deputy, proposed experimenting with both methods.

Political channels and blocking effects

On October 28, Marine Le Pen and Gérald Darmanin each opened a WhatsApp channel, and LFI's channel "L'équipe de Mélenchon officielle" gathered around 155,000 members. Blocking or ignoring content affects the user's feed but does not necessarily impact the collective visibility of the blocked accounts. On X, a blocked account can no longer follow or message the user, and the user stops receiving content from that account. However, the collective effect of blocking remains unmeasured, as platforms do not publish the exact weight of a block in their algorithms.

Impact of blocking on algorithms

André Janones argued that blocking could undermine the strategy, as the algorithm may interpret it as the user not seeing the content. An outraged comment or a mocking share counts as much as a supportive message, as platforms measure reactions rather than their meaning. Ignoring a post deprives it of the fuel that makes it viral. Instagram and TikTok offer a "Not interested" button, which modifies the user's feed but does not necessarily reduce the visibility of the account for others.

Misinformation about Lula's campaign

Thousands of internet users adopted Arkunir's call, believing it to be from Lula's campaign, without verifying what Janones had actually asked in Brazil. Some information sites added "mass blocking" to the list of methods of the Brazilian deputy. Political scientist Gaspard Estrada noted that Janones rather recommended ignoring adverse publications. In Brazil, the orders came from a single command center, while in France, various interpretations were proposed, observes Le Monde.

Petition against CNews

A petition against CNews passed from 45,000 to 200,000 signatures in two days. CNews appears on some block lists, and these days, the channel is the subject of another online mobilization. On October 28, green deputies Benjamin Lucas and Léa Balage El Mariky had asked for the removal of the accreditation of several media from the Bolloré group, including CNews, and relaunched the debate on the channel. However, the signatories should not expect CNews to disappear from the TNT.

Misinformation in live broadcasts

Images shot in Albania, broadcast live on two French channels. The media can also spread fake news. On October 24, CNews, in particular, located porte d'Italie in Paris, images of masked demonstrators, and franceinfo police behind water cannons, announced live from Lyon. The Associated Press agency established that the two sequences had been shot in Tirana, Albania. The presenter of franceinfo corrected three minutes later, and CNews admitted a "error in the control room." On social networks, screenshots had already begun to circulate, and nothing guarantees that the viewers of the sequence saw the correction.

Public trust in social media

According to a study by the Descartes Foundation conducted among 4,000 people aged 16 and over, 43.4% of French people regularly get their information from social networks, but only 17% trust them to get reliable information. When faced with dubious information, 67.5% declare to multiply the sources. These are declared practices, and not a measure of what the respondents actually share.

Regulation of online campaigns

The rules that frame the online campaign before 2027. But safeguards are put in place to best frame the dissemination of political content, and at the dawn of the French presidential election, the government has taken the lead. On October 21, the head of cabinet of the Prime Minister gathered at Matignon the representatives of Meta, Snap, TikTok, Reddit, Discord, and Google. In its statement on October 23, Matignon announced an enhanced cooperation with the platforms, with a single entry point attached to PHAROS, more than a hundred reports, and suspensions of accounts for half of them.

Controversy over content moderation

On October 24, internet users accused the government of "censorship," noting the low visibility of their demonstration videos. That day, PHAROS removed 112 messages calling for violence, according to RMC, and the volume of publications on the mobilization decreased by half compared to October 17, according to Visibrain. The processions remained calm for the most part, and nothing indicates, according to Le Monde, that the platforms voluntarily reduced the dissemination of images. In a letter to the operators, Matignon also asked them to be ready to cut four social networks, including TikTok and X. No cuts took place, but the government obtained a new tool.

Legislative changes and risks

In the night of October 27 to 28, the Senate adopted by 219 votes against 115 its amendment to the "casseurs-payeurs" bill. The administration could require platforms, via PHAROS, to remove content calling for violence or insurrection within 24 hours, under threat of blocking or de-referencing, without a prior court decision. For the presidential election, the very large platforms must, in accordance with the DSA, evaluate and reduce the risks of their services for elections. Despite this, campaigns of large scale escape their moderation. Le Monde revealed on October 29 an operation of three years against Qatar, made up of racist and anti-Islamic videos, viewed at least 830 million times on YouTube, TikTok, and Facebook, about 120 million in France. The researchers behind the study estimate its budget to be more than 10 million euros, without having established who funded it. The "casseurs-payeurs" text now has to go before the National Assembly, which can still modify it before the presidential campaign.