Anthropic published a blog post outlining its official position on open-weight artificial intelligence models, with CEO Dario Amodei stating the company has never advocated for a blanket ban on the technology. Instead of restricting model weights, the policy paper calls for mandatory safety testing of all sufficiently capable models, tighter controls on the export of advanced semiconductor chips to China, and a crackdown on industrial-scale model distillation.

The publication triggered widespread skepticism across the industry, as it is the only major lab to decline a recent Nvidia-backed letter defending open weights. An Anthropic researcher publicly replied to the post to voice disagreement and thank colleagues who pushed for open weights, while multiple observers argued the focus on biological weapon risks and distillation serves as a rationale to protect the competitive advantage of closed, top-tier models.