Anthropic says it disrupted several potential plots this year by scientists using its models for research that could have helped develop biological weapons (Dustin Volz/New York Times)

Anthropic’s threat-intelligence report says the company halted multiple projects in which scientists used Claude to explore biological-weapon research, shutting down the work after being unable to verify its intent. The report also details Chinese firms Moonshot and DeepSeek employing thousands of fake accounts and millions of real user queries through “transfer stations” to distill Claude’s capabilities for their own models.