Benjamin Guinaudeau (@benguinaudeau) | teapawt

Tweets
All Posts
Sort by
Most recent Popular
Tweets & Replies
Reposts
Media
Search

Benjamin Guinaudeau retweeted

@LeHuffPostNov 3

TikTok attise la violence de l’automne social ? L’étude de ces 16 milliards de vues de vidéos dit autre chose huffingtonpost.fr/france/art…

TikTok attise la violence de l’automne social ? L’étude de ces 16 milliards de vues de vidéos dit...

TikTok attise la violence de l’automne social ? L’étude de ces 16 milliards de vues de vidéos dit...

Une étude bat en brèche les poncifs sur le rôle des réseaux sociaux dans les violences, au moment où certains défendent la possibilité de « bloquer » une partie des plateformes.

huffingtonpost.fr

2

2,917

Benjamin Guinaudeau retweeted

Alec ن Archambault@AlexArchambaultNov 5

Une étude qu'elle est bien, car débunkant scientifiquement le narratif du Gouvernement "les émeutes urbaines & pillages ont été causées par les réseaux sociaux". Une nouvelle fois corrélation ≠ causalité. Car la réalité est toute autre. #MustRead tiktokpourlascience.fr/pdf/v…

Image from Alec ن Archambault (@AlexArchambault)

Image from Alec ن Archambault (@AlexArchambault)

Image from Alec ن Archambault (@AlexArchambault)

1

33

2,018

Benjamin Guinaudeau@benguinaudeauNov 4

Tonight's midterm analyses will surely claim that social media swung the results, one way or the other. Need something to read until the results come in? We tested one of those claims: did TikTok fuel the violence in last month's French protests?

Image from Benjamin Guinaudeau (@benguinaudeau)

Benjamin Guinaudeau@benguinaudeauNov 4

Last month, French high-school students blockaded their schools, followed by mass marches against police repression, the state of schools and the cost of living. On Oct. 22, the PM asked telecoms to be ready to suspend TikTok. But did TikTok really fuel the violence? 🧵

Image from Benjamin Guinaudeau (@benguinaudeau)

Image from Benjamin Guinaudeau (@benguinaudeau)

6

Benjamin Guinaudeau@benguinaudeauNov 4

Next: public data show what was posted, not what each person saw. To study individual exposure, we've launched a voluntary donation of TikTok watch histories. Help us understand TikTok's role in the protests!

Our public data do not suggest that TikTok catalysed the violence. Help us learn more! A QR code links to tiktokforscience.org: share what you saw on TikTok by donating your watch history. Takes 5 minutes. Secured by the University of Zurich.

ALT Our public data do not suggest that TikTok catalysed the violence. Help us learn more! A QR code links to tiktokforscience.org: share what you saw on TikTok by donating your watch history. Takes 5 minutes. Secured by the University of Zurich.

3

Benjamin Guinaudeau@benguinaudeauNov 4

Read the full study: tiktokforscience.org

Scientific study: did TikTok fuel the violence of autumn 2026 in France?

Scientific study: did TikTok fuel the violence of autumn 2026 in France?

Help us understand the mobilisations on TikTok by sharing what you saw.

tiktokforscience.org

1

3

Benjamin Guinaudeau@benguinaudeauNov 4

This calls into question the case for blocking social media. Suspending TikTok would hamper the spread of protesters' demands, while violent footage would keep circulating through media outlets' other channels.

1

2

Benjamin Guinaudeau@benguinaudeauNov 4

Bottom line: we find little evidence that TikTok fueled violence through contagion during this autumn's protests. For demonstrators, it was as much a place for humour and joy as for voicing demands.

1

2

Benjamin Guinaudeau@benguinaudeauNov 4

The split wasn't only between types of producers. Among the media shown, outlets with left-leaning audiences focused more on police violence, those with right-leaning audiences more on looting.

Chart: Left and right are not talking about the same violence. For each outlet, the gap in percentage points between its share of videos on looting and its share on police violence. Outlets ordered from left-leaning to right-leaning audience: Le Media -33, Le Monde -15, Quotidien -5, Brut 0, BFMTV +4, TF1 Info +20, Le Figaro +27, CNews +24. Negative values mean more police violence, positive values more looting.

ALT Chart: Left and right are not talking about the same violence. For each outlet, the gap in percentage points between its share of videos on looting and its share on police violence. Outlets ordered from left-leaning to right-leaning audience: Le Media -33, Le Monde -15, Quotidien -5, Brut 0, BFMTV +4, TF1 Info +20, Le Figaro +27, CNews +24. Negative values mean more police violence, positive values more looting.

1

2

Benjamin Guinaudeau@benguinaudeauNov 4

Information production was split into two sharply contrasting discourses: small creators posted joyful, funny videos, while fear was the dominant emotion in media content.

Chart: Video tone, humour from small creators, fear from the media. Share of each account type's views by emotion expressed, small creators vs media: humour 41% vs 4%; joy, pride, hope 33% vs 13%; anger 22% vs 27%; fear, sadness, shock 16% vs 33%.

ALT Chart: Video tone, humour from small creators, fear from the media. Share of each account type's views by emotion expressed, small creators vs media: humour 41% vs 4%; joy, pride, hope 33% vs 13%; anger 22% vs 27%; fear, sadness, shock 16% vs 33%.

1

2

Benjamin Guinaudeau@benguinaudeauNov 4

Among commenters taking a position on violence, 9 out of 10 condemned it, which doesn't suggest contagion on the platform.

Chart: On TikTok, violence is condemned ten times more often than it is celebrated. Out of 100 comments under videos showing violence, 21 condemn violence and 2 celebrate it; the remaining 77 take no position.

ALT Chart: On TikTok, violence is condemned ten times more often than it is celebrated. Out of 100 comments under videos showing violence, 21 condemn violence and 2 celebrate it; the remaining 77 take no position.

1

2

Benjamin Guinaudeau@benguinaudeauNov 4

Compared with the same account's other videos that day, videos showing violence got 40% more views for media outlets, but 34% fewer for small creators.

Chart: Videos showing violence get fewer views for small creators, more for the media. Compared with the same account's other videos on the same day, a video showing violence gets 34% fewer views for small creators (violence costs views) and 40% more views for media outlets (violence brings in views).

ALT Chart: Videos showing violence get fewer views for small creators, more for the media. Compared with the same account's other videos on the same day, a video showing violence gets 34% fewer views for small creators (violence costs views) and 40% more views for media outlets (violence brings in views).

1

3

Benjamin Guinaudeau@benguinaudeauNov 4

Surprisingly, most views of violence went to media outlets. They produced only 1.6% of videos and drew 19% of all views, BUT accounted for 56% of the views on violent videos.

Chart: Images of violence come mainly from the media. Media produced 1.6% of all videos (small creators 87%, others 11%) and drew 19% of all views (small creators 48%, others 33%), but 56% of the views of videos showing violence (small creators 22%, others 22%).

ALT Chart: Images of violence come mainly from the media. Media produced 1.6% of all videos (small creators 87%, others 11%) and drew 19% of all views (small creators 48%, others 33%), but 56% of the views of videos showing violence (small creators 22%, others 22%).

1

4

Benjamin Guinaudeau@benguinaudeauNov 4

Videos showing violence drew a minority of views: 14%. By contrast, videos mentioning at least one protest demand (schools, cost of living, police repression) drew 32% of views.

Chart: Protest demands drew twice as many views as violence. Out of 100 views related to the mobilisations, 14 went to videos showing violence (flame icons) and 32 to videos mentioning protest demands (placard icons).

ALT Chart: Protest demands drew twice as many views as violence. Out of 100 views related to the mobilisations, 14 went to videos showing violence (flame icons) and 32 to videos mentioning protest demands (placard icons).

1

5

Benjamin Guinaudeau@benguinaudeauNov 4

We find almost no evidence that TikTok was used to coordinate violence. Calls to demonstrate circulated widely, but explicit calls for violence were rare and drew few views.

Chart: Calls to demonstrate circulate, but calls for violence are almost invisible. Out of 1,000 views related to the mobilisations, 10% go to calls to demonstrate and 0.1% to explicit calls for violence.

ALT Chart: Calls to demonstrate circulate, but calls for violence are almost invisible. Out of 1,000 views related to the mobilisations, 10% go to calls to demonstrate and 0.1% to explicit calls for violence.

1

5

Benjamin Guinaudeau@benguinaudeauNov 4

In a new study with Isabelle Guinaudeau and Hugo Subtil, we test this claim on nearly one million public TikTok videos. Findings 👇

Title card: Did TikTok fuel the violence during the French October 2026 demonstrations? Analysis based on 964,000 videos and 16 billion views. A line chart of daily views shows two peaks: 1.7 billion views on Oct. 1 (school blockades) and 2.2 billion views on Oct. 17 (demonstrations). Benjamin Guinaudeau, Isabelle Guinaudeau, Hugo Subtil. OBE Note no. 2026-21, CEPREMAP Well-Being Observatory, tiktokforscience.org.

ALT Title card: Did TikTok fuel the violence during the French October 2026 demonstrations? Analysis based on 964,000 videos and 16 billion views. A line chart of daily views shows two peaks: 1.7 billion views on Oct. 1 (school blockades) and 2.2 billion views on Oct. 17 (demonstrations). Benjamin Guinaudeau, Isabelle Guinaudeau, Hugo Subtil. OBE Note no. 2026-21, CEPREMAP Well-Being Observatory, tiktokforscience.org.

1

6

Benjamin Guinaudeau@benguinaudeauNov 4

Last month, French high-school students blockaded their schools, followed by mass marches against police repression, the state of schools and the cost of living. On Oct. 22, the PM asked telecoms to be ready to suspend TikTok. But did TikTok really fuel the violence? 🧵

Image from Benjamin Guinaudeau (@benguinaudeau)

Image from Benjamin Guinaudeau (@benguinaudeau)

1

1

37