Anthropic and Google DeepMind CEOs propose US-led AI coalition
Dario Amodei and Demis Hassabis made the case at a closed-door G7 meeting in Evian. They suggested forming an international alliance to set safety standards and guardrails for AI development.
Amodei stressed the need for structured access to frontier models and chip export curbs to China. The talks touched on AI misuse risks in cyberattacks, bioterrorism, and espionage. Sam Altman of OpenAI also participated, calling for an international forum for independent testing of model capabilities and risks.
The meeting comes on the heels of the US government's export restrictions on Anthropic's latest models, including Fable 5 and Mythos 5.
Related stories
- Two safety researchers leave Anthropic and Google for METR
- D.C. Circuit calls Claude's refusals a supply chain risk
- Trump adviser's memo puts Amodei at the root of EA
- Amodei warns the UN Security Council about AI risk
- Insiders say Anthropic oversold the rogue AI scare
- Ex-Anthropic engineer's exit post passed 170 million views
Comments
No comments yet. Be the first.
Join the conversation
Sign in with Google to leave a comment. Your name and avatar come from your Google profile, and the comment appears after moderation.