Skip to content
This site is deprecated. See the new version.

Key Debates

This section presents structured analyses of major debates in AI safety. For each debate, we present the strongest arguments on multiple sides, identify key cruxes, and track how expert opinion has evolved.

  • Why Alignment Might Be Hard - Arguments for difficulty
  • Why Alignment Might Be Easy - Arguments for tractability
  • Interpretability Sufficient? - Can we rely on understanding models?
  • The Case FOR AI Existential Risk - Strongest arguments for concern
  • The Case AGAINST AI Existential Risk - Strongest counterarguments

Each debate page includes:

  • Steelmanned positions - Best versions of each argument
  • Key cruxes - What would change minds
  • Evidence and citations - Supporting data for each side
  • Expert distribution - Where informed people fall
  • Implications - How resolution would affect priorities