Dario Amodei Clarifies Stance: Supports Open-Weight Models, Fears Chinese AI
Original: Anthropic’s Dario Amodei responds: doesn’t oppose open-weight models, but fears Chinese AI
Why This Matters
The debate over open-weight AI policy directly shapes U.S. tech regulation and the global AI competitive landscape.
Anthropic CEO Dario Amodei stated on Monday that Anthropic has never advocated for a ban on open-weight AI models, clarifying his position after industry speculation. His primary concern is authoritarian governments, particularly China's CCP, achieving military AI superiority over the U.S.
Anthropic founder and CEO Dario Amodei published a blog post on Monday responding to industry rumors that his company supports U.S. government efforts to restrict open-weight AI models, including those from China. He wrote: 'Anthropic has never advocated for a ban on open-weights models,' using his own emphasis.
The clarification followed an open letter published Friday, signed by Nvidia CEO Jensen Huang — in his first post on X — alongside companies including Meta, Microsoft, Hugging Face, and Mistral, urging policymakers against 'premature restrictions' on open-weight models.
Amodei stated that open-weight models without dangerous capabilities are a 'public good,' providing value to businesses, developers, and researchers. However, he distinguished between open-weight model access and the broader threat posed by China. His core concern is that authoritarian governments — with the Chinese Communist Party described as 'most capable' — could develop AI systems more powerful than U.S. counterparts to achieve 'permanent military superiority' or suppress their own citizens.
Amodei also raised fears about AI enabling biological attacks and warned that open-weight models pose heightened risks in such scenarios due to the difficulty of applying guardrails. He cited a UK AI Security Institute report noting that once open-weight models are released, they cannot be withdrawn.
Among his proposed countermeasures: restricting China's access to advanced chips (an existing U.S. policy) and a formal crackdown on distillation, a technique where one AI learns from another by bombarding it with prompts. Amodei also expressed support for international model safety frameworks.