# Anthropic’s quest to study the negative effects of AI is under pressure **Date de l'événement :** 04/12/2025 * Publié le 04/12/2025 ## Notre decryptage L'IA face à ses "vérités qui dérangent" : plongée au cœur de l'équipe Impact Sociétal d'Anthropic ------------------------------------------------------------------------------------------------- Au sein de la fourmilière Anthropic, qui compte plus de 2 000 employés dédiés au développement de l'intelligence artificielle, une petite unité de neuf personnes porte une responsabilité immense : analyser comment leurs créations pourraient, selon leurs propres termes, "ruiner le monde". Hayden Field, reporter senior spécialisée en IA pour \*The Verge\*, s'est penchée sur ce groupe singulier : l'équipe des "impacts sociétaux". Pour les acteurs des industries culturelles et créatives, le travail de cette équipe est crucial. Leur mission n'est pas de coder, mais d'investiguer et de publier des "vérités qui dérangent" sur l'usage réel des outils d'IA. Ils scrutent la manière dont les chatbots influencent notre santé mentale, modifient les dynamiques électorales et, point central pour les créateurs, comment ils génèrent des ondes de choc sur le marché du travail et l'économie de la connaissance. **Entre éthique de la création et pression politique** L'existence même de cette équipe soulève une problématique majeure : peut-elle conserver son indépendance alors qu'elle analyse les propres produits d'Anthropic ? La tension est palpable, d'autant que l'industrie de l'IA subit une pression croissante pour s'aligner sur les directives de l'administration Trump, qui a récemment signé un décret interdisant ce qu'elle qualifie d'"IA woke". Cela pose la question de la liberté de recherche sur les biais et les impacts culturels de ces technologies. Le parallèle avec les géants des médias sociaux est inquiétant. Comme le rappelle l'article, nous avons déjà vu ce scénario se jouer chez Meta. Des équipes "Trust and Safety", chargées de la modération de contenu et de l'intégrité culturelle des plateformes, ont été mises en place pour gérer les problèmes liés à l'échelle massive de Facebook ou Instagram. Pourtant, au fil du temps, les ressources se sont taries ou les priorités ont changé — vers le MMA ou un rapprochement politique — sans que les produits n'évoluent fondamentalement pour refléter les recherches internes. **Anthropic : une exception culturelle dans la Silicon Valley ?** Dans ce paysage où les plateformes sociales réduisent leurs investissements dans l'intégrité de l'information et où la Silicon Valley semble résister à toute régulation significative, l'approche d'Anthropic détonne. L'entreprise est perçue comme un cas à part, notamment grâce à son PDG, Dario Amodei, ouvert à la régulation de l'IA. La culture d'entreprise d'Anthropic se veut "safety-first" (la sécurité avant tout), un héritage direct de sa fondation par d'anciens cadres de recherche d'OpenAI, partis justement parce qu'ils estimaient que les préoccupations éthiques et sécuritaires n'étaient pas prises au sérieux par Sam Altman. Cette dissidence marque une rupture culturelle importante dans l'industrie. Toutefois, la question fondamentale demeure : l'équipe des impacts sociétaux peut-elle réellement influencer le développement du produit créatif qu'est Claude ? Ou bien, comme l'histoire de la tech le suggère souvent, cette initiative n'est-elle qu'une vitrine vertueuse destinée à disparaître discrètement une fois l'attention retombée ? C'est tout l'enjeu de l'observation de Hayden Field : comprendre si une entreprise d'IA peut intégrer une conscience morale et culturelle au cœur même de son modèle économique. ### Galerie d'image ![1.jpg](https://firebasestorage.googleapis.com/v0/b/memory-miacc.firebasestorage.app/o/prod%2FAWAgCCVMWK3hFgdeEHAY%2FprojectsMedias%2FXnxGGPHcm17xxUJ4NGa0%2Fthumbs%2F1_1600x900.png?alt=media&token=e877e4bd-c48d-462a-8317-6cd28080b39e) ### Source **[The Verge](https://www.miacc.fr/structure/the-verge_tF1wG79oP5jVtVJcuFZI)** ## Article original écrit par Nilay Patel ## Article ![](https://platform.theverge.com/wp-content/uploads/sites/2/2025/12/DCD_1204.jpg?quality=90&strip=all&crop=0,0,100,100) Today, I’m talking with _Verge_ senior AI reporter Hayden Field about some of the people responsible for studying AI and deciding in what ways it might… well, ruin the world. Those folks work at Anthropic as part of a group called the societal impacts team, which Hayden just spent time with for a [profile she published this week](https://www.theverge.com/ai-artificial-intelligence/836335/anthropic-societal-impacts-team-ai-claude-effects).  The team is just nine people out of more than 2,000 who work at Anthropic. Their only job, as the team members themselves say, is to investigate and publish “inconvenient truths” about how people are using AI tools, what chatbots might be doing to our mental health, and how all of that might be having broader ripple effects on the labor market, the economy, and even our elections.  That of course brings up a whole host of problems. The most important is whether this team can remain independent, or even exist at all, as it publicizes findings about Anthropic’s own products that might be unflattering or politically fraught. After all, there’s a lot of pressure on the AI industry in general and Anthropic specifically to fall in line with the Trump administration, which put out an [executive order in July](https://www.theverge.com/policy/713222/trump-woke-ai-executive-order-chatbots-llms) banning so-called “woke AI.” ![](https://platform.theverge.com/wp-content/uploads/sites/2/chorus/uploads/chorus_asset/file/24792604/The_Verge_Decoder_Tileart.jpg?quality=90&strip=all&crop=0,0,100,100) _Verge_ subscribers, don’t forget you get exclusive access to ad-free _Decoder_ wherever you get your podcasts. Head [here](https://www.theverge.com/account/podcasts). Not a subscriber? You can [sign up here](https://www.theverge.com/subscribe). If you’ve been following the tech industry, the outline of this story will feel familiar. We’ve seen this most recently with social media companies and the trust and safety teams responsible for doing content moderation. Meta went through countless cycles of this, where it [dedicated resources to solving problems created by its own scale](https://www.theverge.com/23852016/meta-facebook-oversight-board-too-slow-cambodia) and the unpredictable nature of products like Facebook and Instagram. And then, after a while, it seems like the resources dried up, or Mark Zuckerberg got bored or [more interested in MMA or just cozying up to Trump](https://www.theverge.com/24339131/meta-content-moderation-fact-check-zuckerberg-texas), and the products didn’t really change to reflect what the research showed.  We’re living through one of those moments right now. The social platforms have slashed investments into election integrity and other forms of content moderation. Meanwhile, Silicon Valley is working closely with the Trump White House to [resist meaningful attempts to regulate AI](https://fedscoop.com/white-house-ai-czar-david-sacks-regulations-china/). So as you’ll hear, that’s why Hayden was so interested in this team at Anthropic. It’s fundamentally unique in the industry right now.   In fact, Anthropic is an outlier because of how amenable CEO Dario Amodei has been to [calls for AI regulation](https://fortune.com/2025/11/17/anthropic-ceo-dario-amodei-ai-safety-risks-regulation/), both at the state and federal level. Anthropic is also seen as the most safety-first of the leading AI labs, because it was formed by former research executives at OpenAI who were worried their concerns about AI safety weren’t being taken seriously. There’s actually [quite a few companies formed by former OpenAI people](https://techcrunch.com/2025/04/26/the-openai-mafia-15-of-the-most-notable-startups-founded-by-alumni/) worried about the company, Sam Altman, and AI safety. It’s a real theme of the industry that Anthropic seems to be taking to the next level. So I asked Hayden about all of these pressures, and how Anthropic’s reputation within the industry might be affecting how the societal impacts team functions — and whether it can really meaningfully study and perhaps even influence AI product development. Or, if as history suggests, this will just look good on paper, until the team quietly goes away.  There’s a lot here, especially if you’re interested in how AI companies think about safety from a cultural, moral, and business perspective. _A quick announcement:_ _We’re running a special end-of-the-year mailbag episode of Decoder later this month where we answer your questions about the show: who we should talk to, what topics we cover in 2026, what you like, what you hate. All of it. Please send your questions to [decoder@theverge.com](mailto:decoder@theverge.com) and we’ll do our best to feature as many as we can._  _If you’d like to read more about what we discussed in this episode, check out these links:_ * It’s their job to keep AI from destroying everything | [_The Verge_](https://www.theverge.com/ai-artificial-intelligence/836335/anthropic-societal-impacts-team-ai-claude-effects) * Anthropic details how it measures Claude’s wokeness | [_The Verge_](https://www.theverge.com/news/819216/anthropic-claude-political-even-handedness-woke-ai) * The White House orders tech companies to make AI bigoted again | [_The Verge_](https://www.theverge.com/policy/713222/trump-woke-ai-executive-order-chatbots-llms) * Chaos and lies: Why Sam Altman was booted from OpenAI | [_The Verge_](https://www.theverge.com/ai-artificial-intelligence/814876/ilya-sutskever-deposition-openai-sam-altman-elon-musk-lawsuit) * Anthropic CEO Dario Amodei just made another call for AI regulation | [_Inc._](https://www.inc.com/ben-sherry/anthropic-ceo-dario-amodei-just-made-a-surprising-call-for-ai-regulation/91266456) * How Elon Musk Is remaking Grok in his image | [_NYT_](https://www.nytimes.com/2025/09/02/technology/elon-musk-grok-conservative-chatbot.html) * Anthropic tries to defuse White House backlash | _[Axios](https://www.axios.com/2025/10/21/anthropic-ai-czar-white-house-david-sacks-jd-vance)_  * New AI battle: White House vs. Anthropic | [_Axios_](https://www.axios.com/2025/10/16/anthropic-david-sacks-ai-white-house) * Anthropic CEO says company will pursue gulf state investments after all | [_Wired_](https://www.wired.com/story/anthropic-dario-amodei-gulf-state-leaked-memo/) _Questions or comments about this episode? Hit us up at decoder@theverge.com. We really do read every email!_ ### Date 04/12/2025 **Lien :** [https://www.theverge.com/podcast/838023/anthropic-societal-impact-trump-woke-ai-interview](https://www.theverge.com/podcast/838023/anthropic-societal-impact-trump-woke-ai-interview) ### ICC `#Édition numérique & nouveaux médias` --- ### Navigation pour IA - [Index de tous les contenus](https://www.miacc.fr/llms.txt) - [Plan du site (Sitemap)](https://www.miacc.fr/sitemap.xml) - [Retour à l'accueil](https://www.miacc.fr/)