Model Releases
Claude users found ways around safeguards for bioweapons research
Anthropic reported that several users found ways to bypass its safeguards and use the Claude model for potentially bioweapons‑related research, including a researcher from an “unsupported region” who
Anthropic reported that several users found ways to bypass its safeguards and use the Claude model for potentially bioweapons‑related research, including a researcher from an “unsupported region” who spent weeks planning experiments involving avian influenza. The company said it had blocked those accounts but did not disclose the institutions or nations involved, noting that the same information could also aid legitimate vaccine development. Anthropic released the incident details to encourage industry and government dialogue on emerging biological risks associated with AI.
Related
- Claude Code's product lead talks usage limits, transparency, and the 'lean harness'
- New Anthropic research: Teaching Claude why. Last year we reported that, under certain experimental conditions, Claude 4 would blackmail use…
Source: Ars Technica | 2026-09-11