Anthropic new report on AI's labor impact. Their own data show no measurable increase in unemployment. Also, they admit their framework is "most useful when effects are ambiguous." Ambiguity, like fear, is a tool for preempting labor organizing.
www.anthropic.com/research/lab...
Malgré les frappes, Careem, Uber Eats et Deliveroo ont continué leurs livraisons au Moyen-Orient, mettant les livreurs en danger. Management algorithmique et précarité font d’eux de la chair à canon numérique, illustrant le lien direct entre guerre et capitalisme. www.wired.me/story/with-m...
A deux doigts de penser que les jedi anti-sith sont les vrais sith. Une BD inspirée en partie de plusieurs strips du maître jedi @tienstiens.bsky.social
IA en perspective 1
IA en perspective 2
IA en perspective 3
Glossaire d'éducation populaire sur l'IA, les problèmes sociaux, les retombées politiques, les alternatives possibles. Conçu pour enseignants, formateurs, travailleurs sociaux et journalistes et quiconque souhaite aller au-delà des discours dominants. iaep.be
India AI Impact Summit is coming to a close. I was interviewed on France24 (English) about data centers, data work, sovereignty-as-a-service in the Global South. youtu.be/-bUBWFSVsoo?... World leaders, CEOs discuss AI governance at India summit: Who owns AI ?
Podcast Radio3Scienza ft my interview (in Italian) about RentAHuman: hype, scam, or vibe coding experiment? My bet: in 2 months, it’ll turn out that all matching workers-tasks is done “by hand” (by humans and not by agentic AI) and we’ll never hear about it again. www.raiplaysound.it/audio/2026/0...
Demain dans les kiosques
Antonio Casilli : « Avec l’IA, Capgemini transforme son plan de licenciement en argument marketing »
➡️ https://l.humanite.fr/6Lx
Oui dans l’entretien justement.
Licenciements massifs justifiés par l'IA : mon interview avec Pierric Marissal vient de paraître dans L'Humanité. www.humanite.fr/social-et-ec...
Ieri sera Le Iene, popolare trasmissione di inchiesta della televisione italiana, ha mandato in onda la prima parte di un servizio di approfondimento sulle condizioni lavorative di chi addestra l’IA. Nel servizio anche la mia analisi, che condivido per raccontare questa realtà sommersa.
Egypt DiPLab report 1
Egypt DiPLab report 2
Egypt DiPLab report 3
Egypt DiPLab report 4
We are kicking off 2026 with a new report! The comprehensive study "Data Work in Egypt: Who Are the Workers Behind Artificial Intelligence?" conducted by the DiPLab research collective, provides the first in-depth examination of Egyptian data workers.
diplab.eu/diplab-relea...
“Leisure has been colonized by work” says Institut polytechnique de Paris' @casilli.bsky.social
Monday’s #F24Debate looks at the future of free time in the age of technology - and asks whether we ever truly clock off.
Tune in at 7.10PM Paris Time at france24.com/en/live
Un témoignage à la première personne de Michael Geoffrey Abuyabo Asia, ancien modérateur de chats "intimes" avec des IA (« AI companions »), depuis les bidonvilles de Nairobi.
Vertigineuses questions sur le travail émotionnel et le consentement. Via @casilli.bsky.social
data-workers.org/michael/
Où @casilli.bsky.social rappelle qu'avec l'IA, le volume de micro-tâches ne diminue pas : "plus le marché de l’intelligence artificielle générative grandit, plus on a besoin de réentraîner les modèles"
www.rfi.fr/fr/afrique/2...
yes, i can assure you this is fraudulent research
Sam Udotong in - 1st Co-founder & CTO at Fireflies.ai 10h • ••• We charged $100/month for an Al that was really just two guys surviving on pizza. We scaled Fireflies to a $1B valuation after 6 failures from our original crypto food delivery idea. Nothing motivates you more than the pressure of living month to month with no safety net. Before I explain how 2 broke guys validated a $1B idea, you need to understand we were couch surfing while desperately chasing our entrepreneurial dreams. An Al notetaker was our last hope after 6 ideas our friends claimed were "genius". The best way to validate your business idea is by becoming the product yourself. We told our customers there's an "Al that'll join a meeting." In reality it was just me and my co-founder calling in to the meeting sitting there silently and taking notes by hand. When customers scheduled a meeting, we'd manually dial in as "Fred from Fireflies.ai" We'd sit there silently, take detailed notes, and send them 10 minutes later. After taking notes for 100+ meetings (and falling asleep in many), we were finally able to make enough money to pay the $750/month rent for a tiny SF living room. That was the point when we said let's stop and automate everything. The best prototype was 2 guys surviving on pizza. Validation before automation saved us from our 7th failure.
Yet another prime case of Potemkin AI. I hate seeing an “AI notetaker” in a zoom call recording everything. In some cases, these notetakers were literally just a guy writing everything down. We too often see vaporous hype leading to the credulous embrace of fraudulent systems.
Il 20 nov al CSOA Forte Prenestino ci sarà la proiezione del documentario “In the Belly of AI” (Nel Ventre dell’Intelligenza Artificiale”, regia di Henri Poulain), basato sulle mie ricerche e di cui sono coautore. Podcast della mia intervista su Radio Onda Rossa.
www.casilli.fr/2025/11/16/p...
Antonio A. Casilli keynote at the Cooperative AI conference, Istanbul Planlama Ajansı.
XIV century mosaic at The Chora (Kariye Camii), Istanbul.
Turkish coffee break.
Cooperative AI participants.
It’s been an incredible week in Istanbul. I presented our documentary “In the Belly of AI” at the Cooperative AI conference, discussed recent DiPLab research at the TÜFRAM seminar at Marmara university, enjoyed time with old & new friends, indulged in my unabashed passion for anything Byzantine.
"Gneugneugneu la neutralité des IA c'est que des outils"
Sam Altman
If Sam Altman claimed something about a past event — say, a date or a historical episode — would you take his word for it, or would you double-check the facts elsewhere? Then why take him at his word when he talks about things that haven’t happened yet?
The EU has let US tech giants run riot. Diluting our data law will only entrench their power | Johnny Ryan and Georg Riekeles
We have these huge conferences in expensive cities with titles that overshoot actual achievements by a light year or two (I don't think many of the workers at the hotel in Chicago had their work "Reimagined", for example)
En Norvège les passagers doivent monter par la porte avant du bus, mais certains passent par la porte dérobée. scandasia.com/norway-revie...
DiPLab seminar Martin tironi 21 Nov 2025, ISC-PIF, Paris, 3pm CET.
Sovereignty, justice, and sustainability meet AI infrastructures in places with long histories of political experimentation and environmental struggles—like Chile. On Nov. 21, 3pm, come to our DiPLab seminar ft. Prof. Martin Tironi, Pontificia Universidad Católica de Chile: diplab.eu/martin-tiron...
Amazon MTurk
Today marks the 20th anniversary of Amazon MTurk, launched on Nov 2, 2005, to pioneer “artificial artificial intelligence”: using 500,000 underpaid workers to annotate, review, imitate AI. What a difference 20 years makes! Today they’re still underpaid but there are tens of millions of them.
Quand il s’agit d’IA, le catastrophisme est une branche du marketing.
Cette critique du bouquin d'Eliezer Yudkowsky, tête de pont des doomers de l'IA, met le doigt où il faut : ces Cassandre ne s'inquiètent pas des effets sociaux politiques ou économiques de la technologie ; ils sont seulement obsédés par le temps et sa fin.
www.newstatesman.com/ideas/2025/1...
Titre Le Monde : "Elon Musk lance Grokipedia, son Wikipédia jugé biaisé idéologiquement"
Sur Grokipedia, la page dédiée au racisme accrédite des thèses issues du courant du racisme scientifique.
Allo Le Monde / l'AFP ?
À la page "racisme", il y a littéralement des thèses justifiant l'existence de races différentes et validant la théorie de différence de QI selon ces fausses races, il faut quoi pour que vous puissiez juger vous-même du biais idéologique du truc ?
Oui.
