Skip to content
EA Crux Project

Framework Models

Framework models provide general-purpose analytical structures for thinking about AI risk. They offer conceptual foundations that inform more specific analyses.

Models in This Category

Section titled “Models in This Category”
  • Carlsmith Six Premises: Joe Carlsmith’s framework decomposing existential AI risk into six key premises
  • Instrumental Convergence Framework: Analysis of convergent instrumental goals in advanced AI
  • Defense in Depth Model: Layered safety approach for AI systems
  • Capability Threshold Model: Framework for when specific capabilities become dangerous