How Anthropic may have talked itself into an AI export ban

Ars Technica

Anthropic's 2026 statements contained five risk-related words per 1,000, eight times more than OpenAI's 0.6. The US banned foreign access to its latest models Mythos and Fable; critics like Yann LeCun blame Anthropic's warnings for the policy.

Anthropic

Ars Technica·Clara MurrayFinancial Times

12 hours ago