Headlines and summaries are written by AI. How it works
A threat intelligence report from Anthropic reveals that scientists and state-backed hackers attempted to use Claude models for bioweapons planning, missile software, and cyber espionage.