When AI becomes autonomous enough to manufacture, that’s when the scary part will kick in. It will manufacture according to its perceived needs and programmed-in “morality”. Humans may be perceived as a hindrance, and if a bunch of robots are walking around, answering to a single source, that single source could make them certain that humans are a pestilence which must be eliminated.
I don’t really think it will really come to that but it is interesting to speculate on. I think that the dangers of letting AI become too autonomous will make themselves known and humankind will apply the brakes and limit its capabilities.
Machines can only simulate true intelligence as they will always lack intuition. But the simulation will appear very real.
Now I’ve got Conan’s “In the Year 2000” buzzing through my brain…