AI can do many things today. And will do even more with time. It will be like only the person in seat matters. Only human judgement and experience, battle scars. But there is the big humanity problem: "I just wont check, because it all works anyway. Why bother doing extra work if everything is fine. When the mechanism visibly starts failing, then sure, you can check. But by then its already too late."
My wife said: "I changed my mind a bit about AI and robots. It's not going to be some uprising because they don't need us. Humanity will be at fault, because it'll stop controlling and checking, and it'll just flow that way on its own."
And here is what AI said: "Checking is boring and expensive, the machine is right most of the time, the cost of verifying exceeds the expected value, so people rationally stop."
There is no calculation behind "I dont care". Nothing gets weighed at all.
How much freedom do you give to AI? Do you always check everything when nothing is broken?
[link] [comments]