Anthropic CEO Dario Amodei Clarifies Stance: No Ban on Open-Weight Models, But Fears Chinese AI

Anthropic founder and CEO Dario Amodei on Monday directly addressed industry speculation that his company supports banning open-weight AI models, stating unequivocally that Anthropic has never advocated for such a ban. In a blog post, Amodei wrote, "Anyone who has read my past writing should know that I don't regard such bans as a useful measure."

The clarification came after Nvidia CEO Jensen Huang posted an open letter on X last Friday, co-signed by companies including Meta, Microsoft, Hugging Face, and Mistral, urging policymakers not to impose "premature restrictions" on open-weight models. The letter did not name China, but the broader industry debate has centered on allegations that Chinese AI labs are advancing through techniques like distillation, where a model is bombarded with prompts to learn from another model.

Open-Weight Models as a Public Good

Amodei drew a clear distinction between open-weight models and the threat posed by China. "Open-weights models that don't have dangerous capabilities are a public good: they don't cost anything besides the compute needed to run them, and they provide value to businesses, developers, and researchers," he wrote. He emphasized that his longstanding fears about AI are not about businesses using open-weight models, even those from China, but about authoritarian governments building more powerful models to achieve "permanent military superiority" or repress their own people. He identified the Chinese Communist Party as the "most capable" such government.

Amodei also warned that open-weight models are more dangerous in scenarios involving biological attacks, not just cybersecurity threats, because it is difficult to apply guardrails or monitor their usage. He cited a UK AI Security Institute report noting that once open-weights are released, they cannot be withdrawn. This contrasts with arguments from open-source advocates who say access to powerful models helps defenders protect themselves.

Actions to Thwart Chinese AI

Amodei listed specific measures he believes would counter China's AI progress:

• Restricting China's access to powerful chips, a longstanding U.S. policy.
• Formally cracking down on distillation, which the U.S. has threatened with sanctions if it involves IP theft.
• Supporting global model safety testing, especially if China agrees to participate.

Amodei expressed optimism about recent moves by the Trump administration and industry proposals to apply safety testing to the most capable models regardless of origin or openness, while exempting less capable models from startups and academia. He noted that effective testing would need to be global, adding, "I think this may actually be possible ... limited cooperation around preventing AI biological weapons may be possible because it is in China's interest too."

Anthropic's position now aligns more closely with that of OpenAI and Google, which have also backed open-weight models, though the debate over Chinese AI capabilities and distillation continues to shape U.S. policy discussions.