Technology · 11 September 2026

Anthropic blocked AI misuse for biological and conventional weapons

AI company Anthropic has stopped attempts to use its models for "malicious activity" that could help develop biological and conventional weapons. Its latest threat intelligence report, published on Thursday, detailed how its Claude models were misused between December 2025 and August 2026. The report included five cases where scientists employed the AI in methods that might aid the creation of biological weapons.

By BriefTea Editorial · 11 September 2026 at 4:01

Reported by BBC · How we write briefs

Get the app

Understand this story

What are biological weapons? What is 'AI misuse'? Who are Anthropic?

More briefs

BusinessLabour MPs urge renationalisation of Royal Mail. BusinessEY partner pay reaches £870k as consulting revenue falls PoliticsUK and EU discuss future free movement as November summit nears GeneralKing Charles won't apologise for slavery on Caribbean tour, palace says TechnologyAI Datacentre Plans Drawn Up Without Green Groups, Probe Finds BusinessFuel Duty to Be Frozen Again in Budget Browse all stories →