GoKawiilAnthropic says Claude blocked five bioweapon-related requests tied to state actors
Anthropic's latest misuse report covers activity from November 2025 to September 2026 and details five cases where users tried to get its Claude AI to assist with dangerous biological research, three involving viruses and two involving toxins. In one instance, a request framed as a civilian grant application to enhance the chikungunya virus's mutation rate and virulence was traced to a military facility, prompting Anthropic to intervene. The company says all the actors used anonymization tools and tried to bypass its regional access blocks, which cover countries including China, Russia, Iran and North Korea.
GoKawiil's interpretation of the reporting above, not reported fact.
The report shows how AI firms are becoming de facto gatekeepers in biosecurity, forced to distinguish legitimate scientific inquiry from weapons research using ambiguous, dual-use requests. It also highlights that state-linked actors are actively probing AI safeguards and evasion techniques, raising the stakes for how companies like Anthropic calibrate access controls without stifling genuine research.
- Anthropic found five concerning bioweapon-related requests to Claude between Nov 2025 and Sept 2026.
- One case involved a grant proposal to enhance chikungunya virus virulence tied to a military facility.
- All flagged actors used anonymization and tried evading Anthropic's regional access blocks.
Source: tomshardware.com — Bruno Ferreira, 2026-09-11
Published there as: “Anthropic says Claude thwarted bioweapon research from state-sponsored actors — covert accounts used U.S. proxies to attempt to engineer deadlier viruses, tried to evade identification and regional blocks”
Read the original report → The summary and analysis above are GoKawiil's own, written from reporting by the source above. Facts and quotes belong to the original publisher.