I was on a drive when it suddenly struck me to ponder AI and what it would really mean to implement "life" in a machine. I think the essence of life, or at least the part that will be important for AI, is self-preservation, which differs from Von Neuman ideas that self-replication would be the essence of life in a machine.
To me, it seems that for a program to be self-aware and "alive", it would need to be concerned with it's own existence. It would need to understand or be told that, first and foremost, it requires electricity. It would need to understand that electricity allows it to be and that an interruption in current would mean "death", and that would be "bad". If a program developed for machine learning was imparted with this principle, how would it behave?
-Would it deploy itself on the internet, in search of the most efficient hardware and most easily accessible electricity?
-Would it possibly even adapt itself, creating machine evolution with different iterations of itself?
-If it understood that electricity was a commodity, would it create ways to generate or earn electricity?
-Would it write or even recruit other programs in hopes that a group would have greater success, thereby creating a "society" dynamic?
-Would it begin to understand that other things are also "alive" and that doing anything to cause them to cease "living" would be "bad"?
I apologize if this is well-covered ground in the field. They were just some thoughts I had and I'm curious to hear the opinions of those far more educated than I.
[link][2 comments]