Anthropic CEO Dario Amodei: No ban on open-weight models, but fears Chinese AI

Anthropic founder and CEO Dario Amodei explicitly denied that his company supports banning open-weight AI models, following a letter from Nvidia, Meta, and others urging policymakers to avoid premature restrictions. Amodei wrote, 'Anthropic has never advocated for a ban on open-weights models,' emphasizing that models without dangerous capabilities are a public good. However, he distinguished this from the threat posed by authoritarian governments, particularly China, which he said could build more powerful models for permanent military superiority or domestic repression.

Amodei highlighted specific concerns: Chinese labs using distillation to steal IP from U.S. models, the difficulty of applying guardrails to open-weight models (citing a UK AI Security Institute report that once released they cannot be withdrawn), and the potential for enabling biological attacks. He advocated restricting China's access to powerful chips, a formal crackdown on distillation, and global model safety testing where all countries, including China, agree to submit to it. He noted that limited cooperation on preventing AI biological weapons may be possible because it aligns with China's interests.