Definition
Coherent Extrapolated Volition (CEV) is a concept introduced by Eliezer Yudkowsky in the context of AI safety and alignment. It suggests that an advanced AI should not simply obey current human commands, but rather extrapolate what humans would want if they knew more, thought faster, were more the people we wished we were, and integrated more sincerely with each other. The ‘coherent’ part implies resolving contradictions in human values into a consistent utility function, aiming to maximize human flourishing based on our idealized preferences rather than our flawed immediate impulses.
Summary
A proposed goal for artificial intelligence where the system acts according to the refined desires of humanity after idealized reasoning.
Key Concepts
- AI Alignment
- Value Learning
- Idealized Human Preferences
- Utility Function
Use Cases
- Theoretical framework for AGI safety
- Ethical guidelines for superintelligent systems
- Philosophical discussions on machine morality