Anthropic rejects ban on open-weight AI models in U.S. debate

Anthropic Chief Executive Dario Amodei said Monday the company does not support banning open-weight AI models, stepping into a widening U.S. fight over how to regulate systems that users can download and modify. In a blog post, Amodei wrote that Anthropic has "not and are not advocating for a ban" on such models, despite criticism from parts of the tech industry. The clarification came days after Nvidia, Microsoft, Meta and Palantir backed a letter urging policymakers to avoid early restrictions on open-weight models. Amodei said Anthropic agrees those models can broaden access and competition, but disputed claims that they inherently help cyber defenders more than attackers. He argued policymakers should focus instead on keeping advanced chips from authoritarian governments, curbing industrial-scale distillation and requiring safety tests for sufficiently capable models.