Anthropic Blocks AI‑Enabled Biological Weapon Plans

Anthropic says it has identified and disrupted attempts to use its Claude AI model for malicious activity that could support the development of biological weapons.

The firm’s threat‑intelligence report outlines how its technology has been misused for a range of harmful activities, including fake dating apps, hotel Wi‑Fi scams, surveillance of dissidents, and the creation of weapons of mass destruction.

Over the past eight months, Anthropic flagged five incidents that could aid biological weapon development and six that supported conventional weapon systems such as firearms, missiles and armed drones.

Other AI developers—from Google to open‑source projects—have issued similar reports, underscoring a growing industry trend toward transparency and preventive action.

The report prompted calls for a pause on advanced AI research. Senator Bernie Sanders has demanded a halt to AI super‑intelligence, while President Donald Trump defended continued AI progress, warning that a pause could leave the U.S. at a disadvantage.

Anthropic’s head of threat intelligence, Jacob Klein, emphasized that the situation is nuanced: “It’s not about a single villain aiming to kill all humans,” he said. “The risk emerges from many actors seeking to misuse precision language models for harmful ends.”

In line with its commitments, the company has shared relevant intelligence with authorities, industry partners, and has enhanced its monitoring processes to better detect and disrupt future misuse of its models.