6 citations · 17 across the 9 of their papers we have counts for
1 paper · 1 filter
Koen Holtman
Corrigibility is a safety property for artificially intelligent agents. A corrigible agent will not resist attempts by authorized parties to alter the goals and constraints that we…