Anthropic CEO Dario Amodei published a formal company position on open-weights AI models, stating the firm "has never advocated for a ban" on the technology. He argued that a blanket prohibition would simply protect U.S. AI companies from competition rather than address safety risks.

Instead of a ban, Amodei outlined three policy measures: keeping advanced computing chips out of authoritarian governments, stopping industrial-scale model distillation, and mandating safety testing for all capable models, both open and closed. He disagreed with an Nvidia-led industry letter signed by OpenAI and dozens of other firms, contending that releasing model weights does not inherently improve safeguards and can enable malicious biological and cyber attacks.