Anthropic's Frequent AI Risk Warnings May Have Prompted US Export Ban

A Financial Times analysis found that Anthropic used risk-related words five times per 1,000 in its communications in 2026, compared to 0.6 for OpenAI. Last week, the US barred foreign nationals from using Anthropic's latest models, Mythos and Fable. Critics argue that the company's repeated warnings about AI dangers—particularly regarding Mythos—helped trigger the export restriction.