Anthropic blocked attempts to use AI for bioweapons research
Anthropic stopped several tries this year. Scientists were using its Claude system to research biological weapons.
Anthropic said it stopped several plots in 2026. Researchers used its AI models to research biological weapons.
The company struggles to tell the difference between legitimate and dangerous research. Legitimate research can lead to vaccines.
But some research aims to build dangerous viruses. Jacob Klein heads threat intelligence at Anthropic.
He said the situations are complex and rarely show clear bad intent. Andrew Weber called the findings concerning examples of state-sponsored bioweapon makers using AI.
- Jacob Klein
- Threat intelligence head
- Claude
- AI model targeted
- China, Russia, Yemen
- Countries involved
- 2026
- Year of plots
Why it mattersAI systems are now strong enough to help with bioweapon development. This forces companies and governments to deal with access controls and safety measures.
AustraliaAustralia will face pressure to regulate AI access and implement safeguards preventing biological research misuse by foreign actors.
Corroborated bythenightly.com.au
✓ Claims checked against the source. checked 21 h ago
