Anthropic banned foreign-based operator accounts that used Claude AI to generate fabricated news and U.S. midterm coverage supporting Russian, Iranian, and North Korean narratives through a fake news website.