Particle.news

G7 Confronts U.S. Export Controls as AI CEOs Join Evian Talks

Leaders are pressing companies for voluntary safety pledges while urging Washington to address export rules that have cut allies off from Anthropic's most advanced models.

Overview

  • Chief executives from OpenAI, Anthropic and Google DeepMind attended a working lunch with G7 leaders in Evian on Wednesday to discuss model access, frontier cyber risks and youth safety.
  • The U.S. Commerce Department imposed export controls that barred non‑Americans from using Anthropic's Fable 5 and Mythos 5, forcing Anthropic to suspend access to those models for many customers.
  • European officials publicly urged Washington to reconsider what they called discriminatory measures and pushed for 'tech sovereignty' plans to build local compute and model capacity in the EU and Canada.
  • Frontier AI firms are expected to offer a package of voluntary commitments on youth protection and high‑risk cyber and biological misuse that leaders hope will serve as an interim governance framework.
  • The dispute elevated private AI companies into a diplomatic role and could prompt allies to invest in alternative models and infrastructure, which would reshape who controls advanced AI access and tools for businesses and researchers.