signet : Can A.I. “Go Rogue”? — Joshua Rothman · The New Yorker
langage
capitalisme
intelligence artificielle
Joshua Rothman signe un petit papier salutaire de clarification sur la réalité derrière les discours des tenants de l'IA (cartel) et la prolifération de paniques dans le reste de la presse. C'est en anglais, mais tout de même mieux que ma bafouille.
(...) The version you’ve just read relies heavily on anthropomorphic language—that is, it employs words that are typically used to describe people. It presents the agents as individuals, acting independently and under their own initiative, while having thoughts, hatching plans, and making decisions. Yet this isn’t the only way of describing what happened: it’s equally possible to talk about the agents as mindless programs, blindly following instructions they don’t understand. They are, after all, just bits in a computer.
As anyone who uses A.I. knows, models are both very smart and very dumb; they are shockingly knowledgeable, but also out of touch with reality and many human norms. It’s up to users to steer them—and yet, in this case, the user had been replaced by a harness, which is just a piece of software. “Adding powerful computer hacking tools to a harness, and then allowing it to run an LLM-powered Ask → Act → Report for days on end, with no attempt to monitor what it’s up to, is spectacularly negligent,” Newport concludes—like “strapping a weedwhacker to your dog to see if it will end up cleaning the overgrowth in your backyard.” If that plan were to go awry, you’d be laughed at for saying that your dog-weedwhacker “agent” had “gone rogue.” The obvious truth was that you’d simply decided to unleash chaos.
Il convoque notamment Cal Newport qui est peut-être plus connu pour sa production sur la productivité que pour sa carrière de chercheur en informatique. Il semble avoir flairé le potentiel en audience et semble dorénavant alterner entre vidéos sur les systèmes d'organisation du travail et décryptage du discours de Big AI sur sa chaîne YouTube.
Dans tous les cas, c'est une bonne remise à zéro du palais sur les enjeux de pouvoir quant aux mots qu'on utilise pour parler de tout cela. L'auteur réussit à le faire sans s'énerver, ou faire du sarcasme pour compenser, ce qui mérite franchement une médaille.
👋
📣
🕵️
📜
🌈