TechClaude users found ways around safeguards for bioweapons research

Anthropic says it stopped multiple attempts this year by scientists seeking to use its technology for research that could aid biological-weapons development. It cited five cases involving users who circumvented controls or obscured their research aims, including some from Russia, China and Iran, where access to its models is prohibited.

artificial intelligencebiological weaponsAI safety

AArs Technica★★★☆☆2026-09-11 13:02Original