Particle.news
Get it on Google Play
Download on the App Store
19 ARTICLES
·
4w ago

U.S. Orders Anthropic to Block Foreign Access; Company Pulls Its Two Most Powerful Models

The Commerce Department invoked export controls after reports that built‑in safeguards were bypassed, creating a global shutdown that has prompted European calls for greater AI sovereignty.

[Samuel Boivin / NurPhoto / NurPhoto via AFP]
Image
L’application Claude Fable 5 lancée le 9 juin par l’entreprise Anthropic, avant que celle-ci ne soit contrainte par les autorités américaines de la suspendre, trois jours plus tard.
L’entreprise, dont le siège est situé à San Francisco, en Californie (ci-contre), affirme tout mettre en œuvre pour rétablir rapidement ses services d’IA.

Overview

  • Anthropic disabled Fable 5 and Mythos 5 for all users on Friday after receiving a Commerce Department directive to cut access for any foreign national, saying it cannot reliably filter users by nationality.
  • U.S. officials cited national‑security export controls and media reported a reported jailbreak by a user or company that demonstrated a way to bypass safeguards in the models.
  • Anthropic disputes the scale and proportionality of the government’s move, calling the evidence of a limited circumvention a “misunderstanding” and saying other, less sensitive models remain operational.
  • The suspension has provoked urgent political reaction in Europe, where leaders said the episode shows the risks of dependence on U.S. AI platforms and renewed calls to boost European AI capacity and regulatory clarity.
  • The case sharpens longer‑running tensions between Anthropic and the U.S. government—after the Pentagon cut contracts in March—and could affect the company’s commercial plans, regulatory practice for powerful models, and corporate customers who lost access.