Anthropic clarifies its view on open-weights models

Anthropic CEO Dario Amodei said the company does not support a ban on openweights models, responding to recent debate over whether US firms should be barred from using such systems, including some developed in China. In the post, Amodei argued that openweights models can be useful for businesses, developers, and researchers, but said the bigger concerns are national security risks tied to authoritarian governments, cyber and biological misuse, and the difficulty of applying safeguards once model weights are released. He said Anthropic instead supports three measures: restricting powerful chips and chipmaking equipment from reaching China, stopping largescale model distillation efforts, and requiring mandatory safety testing for all sufficiently capable AI models before release.