第 1 則更新，共 1 則最新第 1 天
18 Days, One Order: How a US 'Kill Switch' Cut Europe Off From Frontier AI
The ban on Anthropic's foreign users lasted barely three weeks. The political shock is still unfolding — and Berlin is already rewriting its AI playbook.
最初掌握的資訊
- On June 12, 2026, the U.S. Commerce Department ordered Anthropic to block foreign users from its most advanced AI models.
- Secretary of State Marco Rubio directed U.S. diplomats on July 22 to downplay "kill switch" talk in American technology.
- Germany's digital state secretary Thomas Jarzombek has publicly framed the episode as a European tech-sovereignty problem.
當時尚無定論
- No source in the research file states the ban's exact end date or independently verifies the 18-day duration.
- Whether the Commerce order rested on a specific legal authority, and whether that authority has been challenged in court.
- The full list and technical specifications of the Anthropic models affected — "Fable and Mythos" appear in one subheading , but details sit behind truncated text.
我們說過將關注的
- Whether Germany's procurement tilt toward European AI providers and any data-center expansion translates into binding policy before the next G7 cycle.
- Whether allied governments publicly accept Rubio's "no kill switch" reassurance, or hedge their AI procurement in response.
