Anthropic, the only one absent from 'support for open source'… rebuts, saying it never argued for banning open models

Anthropic faced backlash for being the only major U.S. AI company not to sign a pro-open-weight statement.

CEO Dario Amodei said the company never called for banning open models outright.

He argued the bigger risk is China’s large-scale model distillation and potential military, cyber, and bio misuse of powerful AI.

Amodei pointed to tighter export controls, enforcement against illicit distillation, and mandatory safety testing for frontier models as better responses.

The dispute has widened into a national security debate amid the rise of strong Chinese open-weight models and reported attacks on Hugging Face.