Anthropic CEO Dario Amodei clarifies firm never backed ban on open‑weight AI models FAST-MOVING

Anthropic Dario Amodei open‑weight AI models US‑China AI policy Kimi K3

Anthropic chief executive Dario Amodei stated that the company has never advocated a prohibition of open‑weight AI models, calling such a ban ineffective. He emphasized that open‑weight models without dangerous capabilities are a public good, but warned that authoritarian regimes could use powerful AI for military advantage or repression. Amodei proposed three policy measures: restrict advanced AI chip and equipment exports to authoritarian states such as China, curb industrial‑scale model distillation, and require mandatory safety testing for sufficiently capable AI systems, whether open or closed. The clarification follows a wave of debate in the United States, where officials are weighing restrictions on Chinese open‑weight models such as Moonshot AI’s Kimi K3, a 2.8‑trillion‑parameter model that has drawn attention for its performance. Anthropic also noted it did not sign the open‑weights letter endorsed by 77 technology firms. Chinese developers, meanwhile, report having to use VPNs and other circumvention tools to access foreign AI models despite official rhetoric about openness, and China’s Ministry of Commerce has condemned U.S. actions as “AI hegemony” and warned of counter‑measures. The United States announced plans to investigate and potentially sanction Chinese AI firms over alleged model distillation and intellectual‑property theft.

Entities: Anthropic · China · China · Chinese Ministry of Commerce · Chinese government · Dario Amodei · Kimi K3 · Moonshot AI · Open‑weight AI models · United States · United States · World AI Conference (2026)

Claims

What the coverage asserts, and how well corroborated each claim is across sources.

[○ 1 SOURCE] Xi Jinping delivered a keynote speech at the 2026 World AI Conference in Shanghai, emphasizing open‑win cooperation, risk management, inclusive values and shared responsibility for global AI治理. (b43e998a-ef8d-45ad-a871-c4967415e950)
[○ 1 SOURCE] China's Ministry of Commerce called the U.S. actions a typical AI hegemony and said China would take necessary measures to protect its legitimate rights. (53eca281-cdad-4dc0-a305-212d3d4aa059)
[○ 1 SOURCE] China announced the establishment of the World AI Cooperation Organization, headquartered in Shanghai, with 29 founding member states. (5a53cd6b-235a-41d9-b1a3-ff8b0519770e)
[○ 1 SOURCE] Chinese AI developers report having to use VPNs and other tools to access foreign AI models despite official claims of openness. (existing)
[● 3 SOURCES] Anthropic CEO Dario Amodei stated that Anthropic has never advocated a ban on open‑weight AI models. (177f14b0-3a2b-4f9d-8710-84b7bf908d20)
[● 2 SOURCES] The United States announced plans to investigate and potentially sanction Chinese AI companies, accusing them of model distillation and intellectual‑property theft. (3db24e31-f838-4499-89f6-20788bc4f4a5)
[● 3 SOURCES] Moonshot AI released the Kimi K3 open‑weight model with about 2.8 trillion parameters, the first model in that size range. (existing)
[● 7 SOURCES] Dario Amodei proposed three policy measures: restrict advanced AI chip exports to authoritarian regimes, curb industrial‑scale model distillation, and mandate safety testing for sufficiently capableAI (existing)
[○ 1 SOURCE] Moonshot AI released the Kimi K3 open‑weight model with roughly 2.8 trillion parameters, the first model of its size category. (Model release)
[● 2 SOURCES] China's Ministry of Commerce called the U.S. action 'AI hegemony' and warned it would take necessary counter‑measures. (Chinese ministry response)
[● 2 SOURCES] President Xi Jinping at the 2026 World AI Conference advocated open, cooperative global AI governance and announced support for developing‑country participation. (Conference speech)
[● 8 SOURCES] Anthropic has never advocated for a ban on open‑weight AI models. (existing)