Anthropic has rejected claims that it supports banning open-weight AI models, instead calling for mandatory safety testing for all frontier AI systems amid intensifying US policy debates over AI security and China.
Anthropic CEO Dario Amodei has rejected claims that the company supports banning open-weight AI models, instead urging mandatory safety testing for all frontier AI systems, whether open-weight or closed-weight, before public release.
“Let me state it clearly so that there is no doubt: Anthropic has never advocated for a ban on open-weights models,” Amodei wrote in a blog post. He maintained, however, that open-weight models pose greater security risks because it is “difficult to apply guardrails” or “monitor their usage.”
Amodei also called for stronger US measures to slow China’s AI progress, including tighter AI chip export controls and action against alleged model distillation.
The debate has intensified following Chinese startup Moonshot AI’s release of the open-weight Kimi K3 model, which has challenged leading US AI systems. Meanwhile, OpenAI, Nvidia and Microsoft are among the companies urging Washington not to restrict open-source or open-weight AI models. Anthropic did not join that coalition, prompting criticism that it favours regulations benefiting closed-model providers, an allegation Amodei denied.
Nvidia CEO Jensen Huang continues to advocate for open AI models while meeting US lawmakers on AI policy. Senator Mark Warner has also backed mandatory safety testing for frontier AI systems.
The policy debate comes as the Trump administration prepares a framework for reviewing advanced AI models. Proposals under discussion include creating an independent regulator to assess frontier AI systems before deployment, while allegations of model distillation by Chinese AI companies continue to shape US-China AI policy.
