Anthropic’s Frequent AI Risk Warnings May Have Led to US Export Ban, Critics Say

A Financial Times analysis found that Anthropic used risk-related words five times per 1,000 in its communications this year, compared to just 0.6 for rival OpenAI. Critics argue that the company’s repeated warnings about advanced AI dangers, particularly regarding its Mythos model, contributed to Washington’s decision to bar foreign nationals from using Anthropic’s latest models, Mythos and Fable. The ban has sparked political debate, with some technologists blaming Anthropic’s own rhetoric for triggering the export restrictions.