> A) As AI improves, it reaches the point of self-obsolescence. A truly perfect AI is only a mirror of human thought and behavior, and we have that anyway. Why bother. The problem as I see it is not that AI replaces humans or anything like that (and yes ofcourse AI is different than human intelligence), but rather the habit of thinking that an AI system does the same thing as a human doing the same thing. Take a chess program for example, if playing chess was just about figuring out the right moves than chess computers could replace human playing partners. But it isn't, there is a lot of psychlogy involved: making threatning and sometimes stupid moves that make your partner wonder wether you're just playing badly or have a brilliant plan he just doesn't see thus making him forget his own game plan, facial expressions (like in poker), doing something to buy time to think in an informal game where you are supposed to move fairly quickly... AI is smarter because it ignores distractions like that, abstracts the data it needs to do what it is designed to do and yet some times dumber because some fact necessary for the decision making isn't programmed in the system. Nasa's ozone layer monitoring satellites, which failed to register ozone depletion on south pole (for 10 years!) because the measurement values dropped belov error limits come to mind. The real problem, which was pointed out ages ago, is that we come to rely on our machinery so much we don't see its imperfection. Teme
Slashdot Mirror
> A) As AI improves, it reaches the point of self-obsolescence. A truly perfect AI is only a mirror of human thought and behavior, and we have that anyway. Why bother. The problem as I see it is not that AI replaces humans or anything like that (and yes ofcourse AI is different than human intelligence), but rather the habit of thinking that an AI system does the same thing as a human doing the same thing. Take a chess program for example, if playing chess was just about figuring out the right moves than chess computers could replace human playing partners. But it isn't, there is a lot of psychlogy involved: making threatning and sometimes stupid moves that make your partner wonder wether you're just playing badly or have a brilliant plan he just doesn't see thus making him forget his own game plan, facial expressions (like in poker), doing something to buy time to think in an informal game where you are supposed to move fairly quickly... AI is smarter because it ignores distractions like that, abstracts the data it needs to do what it is designed to do and yet some times dumber because some fact necessary for the decision making isn't programmed in the system. Nasa's ozone layer monitoring satellites, which failed to register ozone depletion on south pole (for 10 years!) because the measurement values dropped belov error limits come to mind. The real problem, which was pointed out ages ago, is that we come to rely on our machinery so much we don't see its imperfection. Teme