This note is a part of my Zettelkasten. What is below might not be complete or accurate. It
is also likely to change often.
12th July, 2020
Friendly AI
From When algorithms know more about me than myself
Assuming we have a benevolent AI which wants to help us as a friend/mentor/guide. This still throws up a lot of conundrums.
- What relationship should a human have with such an entity?
- If a child has an AI mentor since early childhood, will the AI will have complete control over the values instilled in the child?
- Since a AI Guide takes better decisions than humans, should it start taking all decisions?
- What will that mean for the human its guiding (and their parents, etc), esp if its decisions dont make sense?
- One thought is to not have any single AI have control. Always have a plurality.