Anthropic Says It Never Called for Ban on Open‑Weight AI Models, Pushes Export Controls and Safety Tests

Key Points

• Anthropic co‑founder Dario Amodei says the company never advocated for banning open‑weight AI models.
• Amodei supports three measures: stricter chip export controls, a crackdown on industrial‑scale model distillation, and mandatory safety testing for capable models.
• The clarification follows an open letter signed by Nvidia’s Jensen Huang and reports that U.S. officials are considering bans on Chinese open‑weight models.
• Anthropic accused Alibaba’s Qwen lab of a massive distillation campaign against its Claude model, using thousands of fake accounts.
• Amodei stresses policy intervention over reliance on individual company enforcement to curb illicit distillation.
• The company views open‑weight models without dangerous capabilities as a public good, but warns against unchecked proliferation.