Safety Models

Safety models analyze the state of AI safety research, practices, and tradeoffs. They help understand what drives safety work and how to improve it.

Models in This Category

• Safety-Capability Tradeoff: Analyzing tradeoffs between safety and capability
• Safety Culture Equilibrium: Dynamics of safety culture within AI labs
• Safety Researcher Gap: Gap between safety researcher supply and demand
• Capabilities-to-Safety Pipeline: How capabilities researchers transition to safety
• Alignment Robustness Trajectory: Trajectory of alignment robustness as capabilities scale