How much information should we program into an artificially intelligent robotic system? Well perhaps that depends on the tasks, which will be assigned. Indeed the purpose of the system makes sense, so let's say the system is to become an advisor to humans in a leadership role of some sort? Perhaps a system like the USS Enterprise main computer to advise Captain Kirk or perhaps to advise our own US Military central command or a CEO of a multi-national conglomerate?

Should we therefore program the system with all the entire library of congress? The entire Internet archives and all the world's knowledge? How much is too much and how should we go about this? Well, perhaps these new digital library initiatives, which are digitizing every book could be placed in to the computer memory banks? Makes sense right?

And now then, what would a system like this be worth? What would it do when there is conflicting data, as we all know that there are conflicting data sets in Wikipedia, Britanica a

nd even our own annals of history? Then there is an issue of perception? For instance if you were an American then the attack on Pearl Harbor was a cowardess event, but if you were a Japanese then your history shows it was a surprise attack and a decisive battle and victory and the best surprise attacks in modern warfare history. Do you see the problem with programming book knowledge into artificial intelligent systems yet? Consider this in 2006.

Tell Your Friend