Anthropic has used risk-related language five times more frequently than OpenAI this year, according to a Financial Times analysis, as critics argue the company's warnings helped trigger a US ban on foreign access to its newest models.
The FT examined official statements, social media posts, and articles from Anthropic and its CEO Dario Amodei, finding that five out of every 1,000 words related to risk, regulation, or restrictions. For OpenAI and Sam Altman, the figure was just 0.6 words per 1,000 — eight times lower.
Political fallout and export restrictions
Last week, Washington barred foreign nationals from using Anthropic's latest models, Mythos and Fable. Some technologists blame the decision on the $965 billion AI company's repeated warnings about AI's societal risks, particularly concerning Mythos.
Yann LeCun, Meta's former chief AI scientist, said the export ban showed that Amodei's "ridiculous fear-mongering" about AI had paid off. "One reaps what one sows," he wrote on social media.
The dispute has alarmed parts of Europe and Silicon Valley, where executives and officials fear the Trump administration may restrict non-US access to frontier models. It is emerging as an early test of how the US intends to oversee increasingly powerful AI systems.
The FT created lists of terms including "harmful," "dangerous," and "misaligned" and calculated their frequency in statements by each company or its CEO. It also used sentiment analysis to compare the positive and negative tone of communications.