"We have not advocated banning open models," says Anthropic CEO in statement

Anthropic CEO Dario Amodei said he is not calling for a ban on open-weight models.

He urged tighter controls on powerful models, chips, model distillation, and mandatory safety testing for all capable systems.

Amodei partially agreed with recent pro-open-weight statements from NVIDIA, Microsoft, and others, but questioned claims that openness helps defenders more than attackers.

The debate highlights the policy challenge of balancing AI innovation with security and misuse risks.