Anthropic misuse report: blocked bio requests, resellers evade bans, 7 Chinese labs distilling Claude
rohanpaul_ai · x · 2026-09-11
Anthropic's latest AI misuse report details five biological misuse cases (including bird flu research and a chikungunya grant proposal tied to a military research institute) and says it blocked requests to make viruses more dangerous. Account bans didn't stop access — a reseller serving the chikungunya researchers returned within days. Anthropic says it can no longer assure that newer Claude models are too limited to meaningfully aid dangerous bio research, so it restricts a wider range of requests. It also identified unauthorized distillation campaigns involving 7 Chinese labs, and claims DeepSeek and Moonshot forwarded customer requests to Claude, exposing sensitive company data and live credentials to Anthropic without customers' knowledge.
Related event: Anthropic's Threat Report Details Bio-Weapon Plots and Distillation(47 posts)→
More from Safety
- Interactive agent surfaces exempted as "attended" executed 7/7 injection attacks — Federal-Teaching2800 · 2026-09-11
- Commenter calls for stricter regulations on wet labs running frontier RL — mimi10v3 · 2026-09-11
- OpenAI opt-out terms may have a loophole: hidden CoT likely doesn't count as 'Output' — dhruv2038 · 2026-09-11
- Stop The AI Race Campaign Gets Mentioned in Washington Post Coverage — DavidSKrueger · 2026-09-11
- Open-Weight Models Close In on the Frontier; Cures and Pathogen Risks Share One Root — Justin_Halford_ · 2026-09-11
- Stanford professor warns AI labs' evidence-free fearmongering drowns out real risk signals — anshulkundaje · 2026-09-11