Anthropic CEO Dario Amodei clarifies that his company does not advocate for a banning of open-weight AI models.
Anthropic CEO Dario Amodei stated on Monday that his company has never advocated for a ban on open-weights models. This clarification follows a coalition of tech companies, including Nvidia and Microsoft, which urged policymakers to avoid premature restrictions on models that users can download and modify. While some government officials have considered banning these models to manage competition from Chinese startups, Amodei argued that protectionist bans would not address primary national security concerns. Instead, he suggested focusing on keeping powerful chips out of authoritarian hands and requiring safety testing for all capable models. Amodei noted that while open-weight models offer customers greater control and expand access to the AI economy, they do not necessarily favor cyber defenders over attackers. He also highlighted the issues of industrial-scale distillation, a training method where smaller models are built using outputs from larger models, which Anthropic recently identified in a Chinese model family.