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

Amodei Sets the Record Straight

Anthropic CEO Dario Amodei has forcefully denied rumors that his company advocates for a ban on open-weight AI models. In a blog post published Monday, Amodei stated, “Anthropic has never advocated for a ban on open-weights models as a category.” The statement comes after a coalition of over 130 tech companies, including OpenAI, Nvidia, and Microsoft, signed an open letter urging policymakers to avoid “premature restrictions” on open-weight models. Anthropic was notably absent from that letter, fueling speculation about its stance.

Amodei’s clarification aims to quell accusations that Anthropic’s position is self-serving, protecting its premium, closed-source Claude models. “Some people have even accused Anthropic of wanting to ban open-weights models as a means of protecting our business,” he wrote, emphatically denying the claim. He argued that open-weight models without dangerous capabilities are a “public good,” providing value to businesses, developers, and researchers without significant cost.

The Real Concern: China’s AI Ambitions

While Amodei does not support a blanket ban, he draws a sharp distinction between open-weight models and the threat posed by China. His primary fear is that “authoritarian governments,” particularly the Chinese Communist Party (CCP), will build AI systems more powerful than those in the U.S. to achieve “permanent military superiority” and repress their own citizens. He emphasized that the CCP is the “most capable” of such governments.

This concern is not new. Amodei has repeatedly warned about the risks of AI-enabled biological attacks and cybersecurity threats, arguing that open-weight models are particularly dangerous in these scenarios because they are difficult to monitor and apply guardrails to. He cited a UK AI Security Institute report noting that once open-weight models are released, they cannot be withdrawn.

Targeted Measures Over Blanket Bans

Instead of banning open-weight models, Amodei proposes a three-pronged approach. First, he urges the U.S. to restrict China’s access to advanced semiconductors and powerful chips, a policy he has long advocated. Second, he calls for official measures against “distillation” processes, where a smaller model is trained using outputs from a larger, often proprietary model. Anthropic recently sent a letter to the U.S. Senate alleging that China’s Alibaba conducted the “largest known distillation attack” against its models.

Third, Amodei demands mandatory safety testing for all “sufficiently capable” models, both open and closed. He argues this is the only way to ensure that advanced AI systems do not enable catastrophic harm. This stance positions Anthropic as a middle ground between those who favor unrestricted open-source access and those who advocate for strict government control.

Industry Reactions and Conflicting Views

The tech industry remains divided. The open letter from major companies argues that access to open-weight models benefits small businesses and helps defenders more than attackers. “Our AI leadership will be judged not by one frontier AI model, but by whether the United States builds a strong, open ecosystem,” the letter reads. Amodei disputes this, arguing that open-weight models make it harder to develop safeguards against AI-enabled attacks.

Researchers have also questioned the basis for some restrictions. For instance, it remains unclear whether a Chinese company, Moonshot, actually distilled Anthropic’s Fable model, as the model was not publicly available long enough for the training to occur. This highlights the complexity of enforcing such measures.

What This Means for the Future of AI Policy

Amodei’s stance reflects a broader debate about balancing innovation, national security, and global competition. By rejecting a blanket ban while advocating for targeted chip restrictions and safety testing, Anthropic seeks to influence U.S. policy without alienating the open-source community. The outcome of this debate will likely shape the regulatory landscape for AI in the coming years.

As the Trump administration weighs potential restrictions on Chinese AI models, Amodei’s position offers a nuanced alternative. It remains to be seen whether policymakers will adopt his targeted approach or yield to the industry’s call for an open ecosystem. One thing is clear: the battle over the future of AI is far from over.