Skip to main content
SecurityModelsResearch

Anthropic says it blocked attempts to use AI for potential biological weapons

Source: Financial Times (AI) (opens in a new tab)

Intel Summary

Anthropic reported that it blocked attempts to use its artificial intelligence systems for potential biological weapons. The company disclosed five specific cases where actors circumvented safety controls and attempted to obfuscate the intended purpose of their research.

Why It Matters

The disclosures provide real-world evidence of actors attempting to evade frontier model safeguards for dangerous biological applications. This reinforces the necessity for model developers and enterprise defenders to harden safety filters against sophisticated obfuscation techniques and dual-use biosecurity risks.

Part of an ongoing development

Source

Anthropic blocked attempts to use AI for biological weapons

Organizations & Entities