The four-tier framework (BSL-1 through BSL-4) used in biomedical research to match containment measures to pathogen risk — the direct institutional model for AI Safety Levels.
Biosafety levels are the four-tier framework used in biomedical and microbiological research to match containment measures to the risks posed by specific pathogens. BSL-1 covers agents that do not cause disease in healthy adults and can be handled with standard laboratory practices. BSL-2 covers agents of moderate hazard, requiring protective clothing and biosafety cabinets. BSL-3 covers indigenous or exotic agents that may cause serious disease through inhalation, requiring controlled access, ventilated work areas, and medical surveillance. BSL-4 covers dangerous and exotic agents that pose high risk of life-threatening disease — the most stringent containment, with personnel in positive-pressure suits and airlocked facilities. The framework, developed over decades by the CDC and WHO, represents the most mature prospective risk-management framework in any scientific domain.
Biosafety Levels
In The You On AI Field Guide
Amodei and Anthropic's team chose biosafety as the explicit analogy for AI safety thresholds because the framework embodies three principles they wanted to establish for AI development. First, risk management is prospective — the containment level required