How Anthropic’s Cautionary Rhetoric May Have Led to an AI Export Ban

According to a Financial Times analysis, Anthropic used risk-related language five times per 1,000 words in its 2026 public statements, eight times more frequently than rival OpenAI. Last week, Washington barred foreign nationals from accessing Anthropic’s newest models, Mythos and Fable. Some technologists argue that the company’s repeated warnings about AI dangers, particularly concerning Mythos, helped trigger the export ban.