Claim.no.T19YM5

LLMs may already have reached the knowledge needed to create and release biological weapons end-to-end, and current models are approaching the point where they could enable someone with a STEM degree but not a biology degree to produce a bioweapon.

Pessimistic
Share

Original Quote

I am concerned that LLMs are approaching (or may already have reached) the knowledge needed to create and release them end-to-end... We believe that models are likely now approaching the point where, without safeguards, they could be useful in enabling someone with a STEM degree but not specifically a biology degree to go through the whole process of producing a bioweapon.

Outlook

Pessimistic

LLMs enabling non-specialists to produce bioweapons would dramatically increase the risk of mass-casualty biological attacks, threatening millions of lives.

Timeframe
imminentImplied
Confidence
approaching / likely near-term

Potential Data Sources for Validation

Controlled red-team evaluations by AI safety institutes or biosecurity researchers (without publishing results) assessing whether AI models can guide a non-specialist through full bioweapon production process; Anthropic's own biosecurity benchmarks and ASL-3 trigger assessments.

About the Author

Dario Amodei

Dario Amodei is the co-founder and CEO of Anthropic, the AI safety company behind the Claude large language model series, and previously served as vice president of research at OpenAI.

Author Perspective

Across 22 tracked claims, Dario Amodei skews pessimistic.

36% optimistic64% pessimistic
published Mar 6, 2026Pessimistic