Our position on open-weights models
| Source: Anthropic News (community RSS)
Tags: Anthropic, Dario Amodei, open-weights, AI policy, national security, open-source AI, China AI
Anthropic CEO Dario Amodei clarifies the company has never advocated banning open-weights models, reframing its national security concerns around two scenarios: authoritarian governments building superior AI, and open-weights models enabling cyberattacks or bio attacks — neither of which banning US business use would address.
Details
Amid reports that US officials are considering restrictions on Chinese open-weights models and a tech industry letter opposing such bans, Anthropic CEO Dario Amodei published a clarifying post stating the company has never supported banning open-weights models and regards them as 'a public good.' Amodei's two stated nightmare scenarios are: first, an authoritarian government — with the CCP cited as 'the most capable threat' — building AI superior to US systems for military or surveillance use; second, open-weights models (from any country) enabling misuse for cyberattacks or biological attacks. He argues US-business bans address neither: state-level adversaries train models secretly for direct military use, while bad actors bypassing legitimate business channels are not deterred by such restrictions. The post references Amodei's earlier essay 'The Adolescence of Technology' and cites the Intelligence Community's 2026 Annual Threat Assessment. The context matters: tech companies and Anthropic competitors have been publicly positioning around this debate, and this post functions as a clarification of Anthropic's stance while maintaining its overall concern about powerful model proliferation.