Anthropic has detailed efforts to prevent its Claude chatbot from being misused for biological weapons research and other harmful activities, including restricting access by researchers linked to state institutions.
In a report covering the past eight months, the company said distinguishing legitimate biological research from dangerous activity remains challenging because similar techniques can be used to develop vaccines and engineer harmful pathogens.
The company also reported other forms of misuse, including Russian media outlets using Claude to generate propaganda related to Moldova’s elections and attempts to develop conventional weapons such as missiles and drones.
A criminal group, identified as GTG50020, also targeted AI access credentials for extortion, with one intrusion seeking between $1.5 million and $2.5 million.
Anthropic said the findings highlight the difficulty of balancing legitimate AI-assisted research with preventing potentially dangerous applications.
Related Tweet:
Anthropic says scientists used its AI for research that could aid biological weapons development https://t.co/9jgzd0i66w pic.twitter.com/FUIqRmXycL
— New York Post (@nypost) September 10, 2026
Also Read:
TIPP Staff
