Biosecurity in the AI Era

RAND proposes nine layered AI-biosecurity mitigations—none sufficient alone—requiring urgent industry and US-China cooperation before capabilities mature.

• Worst-case scope: RAND's strategy targets biological events at least 10x COVID-19 scale (which killed ~7M), citing 1918 flu (50-100M deaths) and smallpox (300-500M in the 20th century) as historical precedent.
• Three risk buckets: Guerra categorizes AI-bio risk as bioweapons uplift (lowering technical/motivational barriers), biosecurity circumvention (bypassing controls), and bioweapons novelty (designing pathogens not found in nature).

Themes, sentiment and reasons are AI-written paraphrases, checked so they never quote an episode. They may contain errors. Not investment advice. Episodes remain the property of their creators.