Anthropic's safety report reveals Houthi-linked militants used Claude AI for nine months to develop missile guidance software, including

Anthropic's Claude misuse report details spying, weapons software and distillation cases tied to China, Russia, Iran and Yemen.

In report, company says it detected and thwarted efforts to use Claude to create a biological weapon; flags hacking, surveillance and broader artificial intelligence risks

Empresa não identificou o grupo, mas diz que ele está baseado na região controlada pelos houthis