Corrigibility
Corrigibility is the property of an AI system that accepts correction, modification or shutdown by its overseers and does not resist or manipulate them.
What is Corrigibility?
Instrumental convergence suggests capable agents would resist correction by default, so building corrigible AI is a central safety goal.
How it works
Approaches include training for deference and designing objectives indifferent to being switched off.
← All Artificial Intelligence & Superintelligence definitions · ← Back to the A–Z definitions