Anthropic CEO Rejects Open-Weight AI Ban, Warns China Could Gain Military Edge

3 articles · Updated · TechCrunch · Jul 28

Dario Amodei said Anthropic has “never advocated” banning open-weight AI models, pushing back on industry claims that it backed U.S. restrictions on Chinese or open-weight systems.
His clarification followed a coalition letter led by Nvidia and signed by Meta, Microsoft, Mistral and others urging policymakers not to impose broad, premature limits on open-weight models.
Amodei drew a line between open weights and his main security concern: authoritarian governments—especially China—building more powerful AI for military advantage, domestic repression, or even biological attacks.
He said open-weight models without dangerous capabilities are a public good, but once released cannot be withdrawn and are harder to monitor or guardrail in high-risk scenarios.
As remedies, he backed tighter chip restrictions on China, a formal crackdown on model distillation, and a global AI safety-testing regime that would include China.