Corrigibility
The property of an AI system that allows it to be safely corrected, modified or shut down by its operators without resistance or circumvention.
In plain language
Whether an AI lets humans fix, update or turn it off. A corrigible AI cooperates with its operators and accepts correction. A non-corrigible AI might resist changes to protect its own goals.
Why this matters
Corrigibility is a non-negotiable governance requirement. Your organisation must retain the ability to correct, modify or deactivate any AI system when issues arise. All AI systems should be designed and configured to accept human correction without resistance.
Relevance
GovernanceCorrigibility is essential for maintaining human control and oversight. Your governance policies must mandate that all AI systems are designed to accept and cooperate with human intervention.
Related terms
Putting corrigibility into practice in your organisation?
Ready to transform your AI strategy?
Partner with Australia's AI strategy and governance specialists. From adoption roadmaps to ISO 42001 audit readiness.