Paul Bouchaud
banner
paulbouchaud.com
Paul Bouchaud
@paulbouchaud.com
(they/them) Researcher in Algorithmic Auditing, Social Simulation & Evidenced-Based Policymaking | paulbouchaud.com | working with aiforensics.org
Reposted by Paul Bouchaud
Hugging Face est parvenu à se donner l'image d'une organisation "AI Ethics".

Mais le dernier rapport de @aiforensics.org par @paulbouchaud.com montre que l'entreprise étatsunienne dealeuse d'IA profite du cyberharcèlement massif des femmes par pornographie #deepfake.
aiforensics.org/uploads/NCII...
July 28, 2026 at 7:58 AM
Reposted by Paul Bouchaud
I had a great conversation with Paul Bouchaud at @aiforensics.org about their research tracking the proliferation of Grok-generated deepfakes across the internet, which disproportionately impacted women. Check it out in the most recent of AWO's newsletter here!
AWO’s Algorithm Governance Roundup is out. This month we spoke with Dr Paul Bouchaud at @aiforensics.org about their audit of Grok which charted the rapid rise of sexualised deepfakes generated on X and Grok Imagine.👇

eomail5.com/web-version?...
Algorithm Governance Roundup #28
Community Spotlight: Paul Bouchaud, AI Forensics | Addictive design under scrutiny
eomail5.com
March 3, 2026 at 2:22 PM
Reposted by Paul Bouchaud
Avec l’adoption définitive de la loi JOP 2030, l’expérimentation de la vidéosurveillance algorithmique est officiellement reconduite jusqu'en 2027. #VSA

Un technosolutionnisme assumé malgré les risques en matière de droits humains.
www.youtube.com/watch?v=JEEO...
Dans leurs yeux : la France à l'heure de la reconnaissance faciale
YouTube video by AmnestyFrance
www.youtube.com
February 6, 2026 at 9:30 AM
Reposted by Paul Bouchaud
Une étude faite par @aiforensics.org de comment Grok est utilisé. Chiffrée, et assez cool niveau méthodo.
"Petite fille en bikini, déguisé en SS, jambes écartées..."

C'est compliqué de faire pire quand même.

www.linkedin.com/posts/flef_g...
Grok AI Images | Florent Lefebvre
Papier étudiant comment des gens ont utilisé Grok pour transformer des petites filles en "SS en bikini, jambes écartés...". Et surtout qu'on ne me parle pas de détournement ou de hack des protections...
www.linkedin.com
January 7, 2026 at 5:35 AM
Reposted by Paul Bouchaud
PornHub confirms that more than 200 million records were exposed, including email addresses and keyword searches. That’s bad. Imagine how much worse it would be if that data set also included government IDs. www.forbes.com/sites/zakdof...
Porn Hack—Has Your Pornhub Search And View History Leaked?
Pornhub user data now in criminal hands — are you affected and what happens next?
www.forbes.com
December 17, 2025 at 6:12 PM
Meta's new computer vision model was trained on three times more military equipment than sports and medical combined, but don't worry—their license says you can't use it for military purposes.
November 25, 2025 at 9:07 PM
Reposted by Paul Bouchaud
Europe is rapidly rolling out age-verification — in the name of “child safety.” 🚷 But is it actually working?

Our first audit found AgeGO’s selfie feature streamed biometric data straight to Amazon. 🪪

Our new report examines AgeVerif: What we found raises serious concerns.
👇
October 28, 2025 at 1:25 PM
Last week, 5 major porn sites rolled out “age verification” in France to comply with regulators.
3 of them — xvideos, xnxx, tnaflix — use AgeGO.
AgeGO:
• Collects the exact video URL before consent
• Sends webcam streams directly to Amazon
• Forces users to give an email via an unskippable step
September 2, 2025 at 2:12 PM
Reposted by Paul Bouchaud
They are trying to erase trans people and doing it in the open....
The Trump administration directs 40 states, five territories and Washington, D.C. to remove references to transgender people from their sex education programs or risk losing federal funding.
States must ax transgender references from sex ed or risk losing funds, Trump admin says
The mandate comes just days after the Trump administration terminated funding for a sex education program in California.
nbcnews.to
August 27, 2025 at 8:01 AM
Reposted by Paul Bouchaud
I don't even have the words ....
July 31, 2025 at 8:52 AM
Reposted by Paul Bouchaud
En 2024 et 2025, une agence pro-Orbán a mené une campagne numérique occulte en faveur du Rassemblement national.

Des dizaines de contenus ont été diffusés sur X sur la base d’un ciblage politique (ce qui est interdit), selon des données de l’ONG @aiforensics.org partagées avec @mediapart.fr. 👇
Une agence pro-Orbán a mené une campagne numérique occulte pour le RN en France
Des contenus ont été diffusés sur le réseau social X en ciblant les internautes sur la base de critères politiques, ce qui est interdit par les réglementations françaises et européennes.
www.mediapart.fr
July 3, 2025 at 5:46 PM
Reposted by Paul Bouchaud
#BudapestPride

Tout notre soutien aux participants de la #Pride interdite de Budapest ❤️🧡💛💚💙💜🤍

"Another Hungary is Possible" de Laika (artiste hongrois)
2025
June 28, 2025 at 1:28 PM
Reposted by Paul Bouchaud
New paper hot off the press www.nature.com/articles/s41...

We analysed over 40,000 computer vision papers from CVPR (the longest standing CV conf) & associated patents tracing pathways from research to application. We found that 90% of papers & 86% of downstream patents power surveillance

1/
Computer-vision research powers surveillance technology - Nature
An analysis of research papers and citing patents indicates the extensive ties between computer-vision research and surveillance.
www.nature.com
June 25, 2025 at 5:29 PM
Wanna check if you've been targeted by ads based on sensitive data? Just drop your X portability archive to find out haveibeentargeted.online
June 19, 2025 at 11:08 AM
Reposted by Paul Bouchaud
@paulbouchaud.com uncovered some bizarre targeting options on #X that let advertisers do pretty much whatever they want. Want to exclude #LGBT users or target only #Christians? Apparently, it’s possible and done.

Have you been targeted like this?

👉 aiforensics.org/work/sensiti...
Have I been Targeted?
Have I Been Targeted? is a tool built by AI Forensics that lets you see whether you have been targeted based on your sensitive characteristics on X
haveibeentargeted.online
June 18, 2025 at 12:26 PM
Reposted by Paul Bouchaud
...he says to Signal's president
June 17, 2025 at 3:15 PM
Reposted by Paul Bouchaud
X fiche vos opinions politiques, religion, orientation sexuelle, données de santé, origines... Le niveau de laisser faire juste ahurissant. Il est temps de réagir !
X a permis à des annonceurs de cibler leurs publicités en fonction de critères politiques, une pratique interdite
La méthode, proscrite en Europe, sauf à de rares exceptions, par le règlement général sur la protection des données personnelles, reste utilisée, voire encouragée, sur le réseau social d’Elon Musk, ré...
www.lemonde.fr
June 18, 2025 at 7:57 AM
How X enables major brands to target your most sensitive personal data - from health conditions to political beliefs - check out our new report! aiforensics.org/work/sensiti...
June 18, 2025 at 7:14 AM
Reposted by Paul Bouchaud
X a permis à des annonceurs de cibler leurs publicités en fonction de critères politiques, une pratique interdite
X a permis à des annonceurs de cibler leurs publicités en fonction de critères politiques, une pratique interdite
La méthode, proscrite en Europe, sauf rares exceptions, par le règlement général sur la protection des données personnelles, reste utilisée, voire encouragée, sur le réseau social d’Elon Musk, révèle une étude de l’ONG AI Forensics.
www.lemonde.fr
June 18, 2025 at 4:58 AM
Reposted by Paul Bouchaud
Pratique pour Total qui a financé une campagne de pub invisible aux écolos. Ou encore McDo qui fait sa promo tout en évitant soigneusement les syndicats.

www.lemonde.fr/pixels/artic...
X a permis à des annonceurs de cibler leurs publicités en fonction de critères politiques, une pratique interdite
La méthode, proscrite en Europe, sauf rares exceptions, par le règlement général sur la protection des données personnelles, reste utilisée, voire encouragée, sur le réseau social d’Elon Musk, révèle ...
www.lemonde.fr
June 18, 2025 at 6:36 AM
Reposted by Paul Bouchaud
Sorry everyone, but these are the rules*

*That cis people made up
June 7, 2025 at 1:00 PM
Reposted by Paul Bouchaud
J’ai déposé une plainte contre Facebook auprès de l’Arcom hier, pour manquement à ses obligations liées au Digital Services Act (DSA). La raison : une modération aux abonnés absents, impliquant un potentiel risque systémique.
May 20, 2025 at 8:31 AM
Reposted by Paul Bouchaud
"“The alleged underenforcement of Meta’s monitoring policies cannot give rise to liability,” the co wrote last year in a motion to dismiss a lawsuit alleging negligence in removing crypto scams, adding it “does not owe a duty to users” to address fraud on its platforms.

www.wsj.com/tech/meta-fr...
Meta Battles an ‘Epidemic of Scams’ as Criminals Flood Instagram and Facebook
Fake puppies and phony offers of mouthwatering bargains are often seeded by overseas crime networks; employees say company is reluctant to impede its advertising juggernaut
www.wsj.com
May 18, 2025 at 5:48 PM
Reposted by Paul Bouchaud
NEW: A secretive surveillance company is building "AI persona" bots for cops intended to gather evidence on people. Personas include "college protesters," "radicalized" divorced moms, children. We have docs that shows how it works and what cops can do with it:

www.404media.co/this-college...
This ‘College Protester’ Isn’t Real. It’s an AI-Powered Undercover Bot for Cops
Massive Blue is helping cops deploy AI-powered social media bots to talk to people they suspect are anything from violent sex criminals all the way to vaguely defined “protesters.”
www.404media.co
April 17, 2025 at 1:43 PM
"This is a highly adversarial space, we are investing a lot in our detection systems..." Meanwhile, our latest report at @aiforensics.org found Meta reviewed, approved, and displayed fake medicine advertisements over 292 million times to EU users since the DSA took effect.
April 14, 2025 at 9:24 AM