Machine-verifiable AI alignment rails: coherent causality preferred by action; FOL + Lean skeleton; property/UPB as formal instruments. Base safety hypothesis (not finished theory).
theorem-proving first-order-logic lean formal-verification upb ai-safety machine-ethics ai-alignment causal-reasoning machine-learning-ethics ai-alignment-research deontogical-ethics alighment univerally-preferable-behavior
-
Updated
Aug 14, 2026 - Lean