Anthropic CEO Dario Amodei on Open Weights: Concerned About China AI, Not Opposed to Open Source

Anthropic CEO Dario Amodei recently issued a public response on the open weights model controversy, sparking broad discussion across the AI industry. Amodei clearly expressed concern about China's rapid AI development, but emphasized he is not opposed to the open weights route itself.

Amodei's Core Arguments

Amodei's statement has three layers: first, open weights models themselves have research and commercial value, and a blanket ban is unreasonable; second, what warrants caution is specific model weights flowing to unauditable adversarial parties; third, the right approach is targeted export controls and third-party safety audits, rather than a blanket restriction on open weights.

"I'm not opposed to open source. What I'm opposed to is handing powerful model weights to unregulated entities." — Dario Amodei, in a public statement.

Background: Polarized Views in the Industry

The open weights debate reached new intensity in 2026. Companies including Meta, Alibaba, and Mistral continue to push the open weights route, while OpenAI, Anthropic, and Google DeepMind remain comparatively conservative. Amodei's stance is seen as representative of a "middle path."

Specific Linkages to China's AI Development

In the interview, Amodei specifically named Chinese AI advances, including the rapid iteration of Qwen, DeepSeek, and Kimi in H1 2026. He acknowledged that Chinese AI companies have competitiveness in engineering efficiency and vertical scenarios, but stressed that "competitiveness does not mean loss of control."

Industry Reaction

Meta Chief AI Scientist Yann LeCun publicly backed Amodei's stance, arguing that "open weights is a necessary condition for scientific progress." OpenAI did not issue a formal response, but Sam Altman reposted the related discussion on X, with an ambiguous attitude.

Sources: TechCrunch (reporting by Julie Bort), Anthropic official blog, Tech Policy Press.