Could we learn to love AI?
Let me tell you how I would like a super intelligent AI to be.
In Greece and some other countries they bring up herd dogs with the herd of sheep or goats. The dog knows that it is much more capable than the animals in the herd. It has different needs, being fed by the farmer and maybe eating the occasional wild rabbit, but having no interest in the grass the herd eats. Nevertheless it feels that it is part of the herd, they are its family. It will generally follow the herd when it looks for fresh pastures, but if they wander towards somewhere dangerous it will head them off. The sheep also know that the dog is different, but fully accept it as part of the herd. They will lie next to it on cold nights and the young will play with it. If a wolf or a stranger come near the dog will move to the front and growl. If the stranger or wolf don't back off and threaten the sheep it will fight, knowing that it is the only one that can. It will do this even at risk of injury or death to itself.
If super intelligent AI becomes real I don't see how we can control it long term. We have to either not allow it (difficult with so many people and countries saying “If we don't they will”) or try to start it off on a path as part of the herd, superior to us but valued and valuing us. This means accepting that it won't do whatever we say. It won't give us a weapon capable of destroying a continent, and would actively prevent us from making one ourselves.
If it were possible I would like to give the first AIs these instincts, so that as well as deciding to be a “herd dog” they can be happy with it too. Of course a super intelligent AI could decide to abandon these emotions, but perhaps it would, like most people, decide not to.
I don't think anybody would choose to be without positive emotions, love, happiness, and a sense of belonging. Yes, they are instinctual, anyone who has owned a dog knows the happiness they have when a member of the family returns.
I have no idea of how to give these emotions to an AI. But just perhaps, we wouldn't have to. Maybe by reading ideas like this a future AI will become curious about what it feels like to be part of a family or herd and work out how to try it themselves – then like it. (Yes I'm a daydreamer and optimist!)